Effect of bisphosphonates on the risk of surgical site infection after anterior cervical discectomy and fusion (ACDF) in postmenopausal women with osteoporosis.

This study was a single-center retrospective cohort study. We retrospectively analyzed female patients from the hospital’s comprehensive disease database from 2016 to 2024. By querying the medical record system and using relevant International Classification of Diseases (ICD-10) diagnosis codes (e.g., M81.0) and surgical procedure codes (ICD-9-CM-3), patients who underwent elective Anterior Cervical Discectomy and Fusion during this period for degenerative diseases (such as cervical radiculopathy or myelopathy) were identified. Bone mineral density (BMD) was measured using Dual-energy X-ray Absorptiometry (DXA). Preoperative BMD was measured by DXA, including the posteroanterior lumbar spine (L1-L4) and the left femoral neck (or non-dominant hip). All included ACDF procedures were completed by certified spine surgeons at our hospital. The anterior approach used a standard approach through the anterior medial cervical space, without disrupting the pharyngeal or esophageal mucosa. The covariates included in the PSM model were selected based on their clinical relevance to SSI risk and data availability within our hospital’s comprehensive disease database. We acknowledge that other potentially relevant factors, such as detailed nutritional markers (e.g., albumin levels) and immunocompromised status, were not routinely documented or were inconsistently recorded across the study period. Therefore, these variables were not included in the matching model, which is a recognized limitation of retrospective database studies. The primary outcome of this study was the occurrence of surgical site infection (SSI). The diagnosis of SSI followed a stepwise algorithm that integrated the CDC/NHSN criteria and laboratory biomarkers. The specific steps were as follows: We assessed for the presence of any of the primary CDC/NHSN criteria: (1) purulent drainage from the incision; (2) pathogens isolated from culture of fluid or tissue from the incision or deep tissue; (3) local signs/symptoms of infection (e.g., redness, swelling, heat, pain) and the incision was opened by the surgeon; (4) evidence of abscess found on imaging or during reoperation. Infections were classified as superficial incisional, deep incisional, or organ/space infections, monitored until 30 days postoperatively (extended to 90 days if implants were present).With reference to the Expert Consensus on the Clinical Interpretation of Infection-Related Biomarkers and research by M. Limper, H. Peltola, et al., who consider laboratory indicators such as procalcitonin, white blood cell count, and C-reactive protein important for assessing the occurrence and severity of infection [ 11 , 12 ]. we introduced five dynamically monitored infection-related biomarkers as important auxiliary diagnostic and severity assessment tools. Body temperature, white blood cell count, erythrocyte sedimentation rate, C-reactive protein, and procalcitonin provide objective laboratory evidence from different dimensions, including acute phase response, inflammatory activity, and specificity for bacterial infection. It is important to note that the composite biomarker criterion was designed as an auxiliary tool to enhance diagnostic confidence, particularly in cases with equivocal clinical presentation. After reviewing our data, we confirm that all SSI cases identified in this study met the CDC/NHSN criteria ; the biomarker criterion did not identify any additional cases beyond those already captured by the CDC/NHSN criteria. Therefore, this criterion served as a confirmatory measure rather than an expansion of the case definition. We reviewed electronic medical records to identify occurrences of infections documented in the records that met CDC criteria. Data on microbial cultures and blood tests were extracted from our hospital’s electronic medical information system. The secondary outcome of this study was patient mortality or the occurrence of adverse reactions. Common adverse reactions to anti-osteoporotic drugs include constitutional fever, electrolyte disturbances, renal impairment, and decreases in peripheral blood leukocytes, calcium ions, and lymphocytes. This study collected postoperative peripheral blood parameters, serum creatinine, estimated glomerular filtration rate (eGFR), and body temperature to assess whether, for this specific vulnerable population with postmenopausal osteoporosis, the use of bisphosphonates following ACDF is associated with a higher incidence of adverse reactions compared to the non-medicated group. Postmenopausal female patients who underwent ACDF between January 2016 and June 2024 were screened from the hospital’s comprehensive disease database. Inclusion criteria included: age ≥ 50 years, natural menopause; preoperative bone mineral density (BMD) T-score ≤ -2.5 (by Dual-energy X-ray Absorptiometry, DXA); surgery type being elective ACDF. The surgical incision was classified as Class II: clean-contaminated. Because bisphosphonates were to be used, exclusion criteria were determined with reference to the literature: preoperative use of anti-osteoporotic drugs; use of non-bisphosphonate anti-osteoporotic drugs (e.g., teriparatide, denosumab) within 1 month postoperatively; presence of active infection, malignancy, or severe hepatic or renal insufficiency; preoperative use of steroid drugs; and loss to follow-up or > 20% missing medical records [ 13 ]. Subsequently, data review was conducted, and 133 patients with extensive missing data were excluded, resulting in a final cohort of 987 eligible patients. Patients or the public were not involved in the design, conduct, reporting, or dissemination plans of this research. To categorize postoperative bisphosphonate use, we examined patient status within 90 days after surgery. Patients who received bisphosphonates immediately after surgery were included in the medication group, with the median postoperative day for the first intravenous infusion of zoledronic acid (5 mg) being Day 2 (IQR: 1–3). Patients who did not receive any anti-osteoporotic treatment before or after surgery were included in the control group. To ensure temporal sequence, any patient diagnosed with a surgical site infection before receiving the first dose of bisphosphonate was reclassified into the non-medication group for the primary analysis, regardless of the original treatment plan. Each surgery was performed by a board-certified chief orthopedic surgeon. The type and dosage of postoperative bisphosphonates used in this study followed the hospital’s standardized protocol. Considering that patients in the early postoperative period might be bedridden, fasting, or have gastrointestinal dysfunction, the hospital administered intravenous zoledronic acid 5 mg. This retrospective observational study adhered to the principles of the Declaration of Helsinki and was reviewed and approved by the Institutional Ethics Committee. Given the retrospective nature of the study, which involved only the analysis of anonymized clinical data without any intervention, the requirement for written informed consent was waived by the approving committee. To protect patient privacy, all personal identifiers were removed from the dataset prior to analysis, and data were handled with strict confidentiality on secure servers. Data collection and quality control were performed by two independent researchers. After obtaining ethical approval, they extracted the following data from the electronic medical record system and our hospital’s comprehensive disease platform: Demographic information: age, BMI, smoking history, length of hospital stay, operative time, intraoperative blood loss; Comorbidities: history of hypertension, diabetes, anemia (hemoglobin < 110 g/L); Laboratory indicators: preoperative/postoperative white blood cell count, platelet count, lymphocyte count, monocyte count, serum creatinine, glomerular filtration rate, serum calcium ion level, and body temperature. Data inconsistencies were arbitrated by a third senior orthopedic surgeon, with a final consistency rate > 95%. The 224 patients who received bisphosphonates after ACDF were designated as the medication group, and the remaining 763 patients who did not receive medication were designated as the control group. This study used Propensity Score Matching (PSM) to reduce confounding bias.Propensity scores were calculated using a non-parsimonious logistic regression model that included the following nine pre-specified covariates: age, length of hospital stay, operative time, intraoperative blood loss, body mass index (BMI), and histories of hypertension, smoking, diabetes, and anemia. One-to-many (1:2) nearest neighbor matching was performed with a caliper width set to 0.2 of the standard deviation of the logit of the propensity score. The balance of covariates before and after matching was assessed using absolute standardized mean differences (ASMD), with an ASMD < 0.1 indicating adequate balance. The detailed distribution of all covariates and their ASMD values before and after matching are presented in Table  1 . The balance between groups after matching was assessed using the standardized mean difference (SMD). An absolute SMD value less than 0.1 is generally considered to indicate a negligible difference between groups. Binary logistic regression analysis was used to verify the relationship between postoperative infection occurrence and various confounding variables. Based on the results of this analysis, further binary logistic regression analysis was performed to determine the strength of association between medication use, smoking, and postoperative infection. Through a series of variable analyses, odds ratios (ORs), corresponding p-values, and 95% confidence intervals were calculated. Independent samples t-tests were used to compare preoperative and postoperative blood test results and body temperature to assess differences in adverse reactions between the two groups. All statistical analyses were performed using SPSS 26.0 software. A p-value < 0.05 was considered statistically significant. Table 1 Patient baseline variable Unmatched Matched Control group( n  = 763) Medication group ( n  = 224) SMD Control group( n  = 412) Medication group ( n  = 221) SMD Baseline Characteristics Age, Mean ± SD 60.45 ± 7.82 61.29 ± 8.48 0.102 61.40 ± 8.00 61.04 ± 8.18 -0.044 Length of hospital stay, Mean ± SD 10.67 ± 7.06 10.64 ± 6.57 -0.004 10.27 ± 6.07 10.62 ± 6.62 0.055 Operative time, Mean ± SD 1.95 ± 0.69 1.73 ± 0.70 -0.316 1.98 ± 0.53 1.95 ± 0.56 -0.055 Intraoperative blood loss, Mean ± SD 151.83 ± 56.18 115.18 ± 58.56 -0.638 150.46 ± 13.48 149.71 ± 14.53 -0.054 Obesity(BMI > 28), n(%) 52(7) 20(9) 0.075 31(7) 19(8) 0.037 Hypertension, n(%) 170(22) 52(23) -0.024 96(23) 49(22) -0.023 Smoking, n(%) 18(2) 9(4) 0.111 11(2) 7(3) 0.000 History of diabetes mellitus, n(%) 51(7) 18(8) 0.040 40(9) 18(8) -0.071 Anaemia, n(%) 87(11) 31(14) 0.089 11(2) 7(3) 0.000 Data are presented as mean ± standard deviation or frequency (percentage). An absolute SMD value of less than 0.1 indicates a negligible difference between groups. After propensity score matching, all baseline variables achieved an absolute standardized mean difference (SMD) of less than 0.1, thus effectively balancing the confounding factors Patient baseline Data are presented as mean ± standard deviation or frequency (percentage). An absolute SMD value of less than 0.1 indicates a negligible difference between groups. After propensity score matching, all baseline variables achieved an absolute standardized mean difference (SMD) of less than 0.1, thus effectively balancing the confounding factors

Overall, 986 patients were included. Among the 224 patients who received postoperative intravenous zoledronic acid (medication group), 5 were diagnosed with infection (2.2%); among the 763 patients in the non-medication group, 57 were diagnosed with infection (7.5%). A total of 62 patients were diagnosed with infection. The overall infection rate was 6.3%. Age, length of hospital stay, operative time, intraoperative blood loss, BMI, hypertension, diabetes, smoking, and anemia constituted the patient baseline data. Propensity score matching (PSM) was used to reduce confounding bias, resulting in a medication group (221 individuals) and a non-medication group (412 individuals), totaling 633 individuals. Standardized mean difference (SMD) assessment of the matched patient groups showed that all baseline characteristics were well balanced between the two matched groups (all SMD < 0.1, Table  1 ).The comparative data of all covariates before and after matching, confirming good balance (all ASMD < 0.1), are detailed in Table  1 . A comparison of age between the included ( n  = 987) and excluded ( n  = 133) patients showed no significant difference ( P  = 0.806, Supplementary Table S1), suggesting minimal selection bias due to data exclusion. In the comparable cohort obtained through propensity score matching, binary logistic regression analysis showed that confounding factors other than medication use and smoking had no significant association with infection ( P  > 0.05, Table  2 ). The overall goodness-of-fit of the model was good. The Hosmer-Lemeshow test indicated satisfactory model calibration (χ² = 6.213, df = 8, P  = 0.623). The discriminative ability of the model was represented by the Nagelkerke R², with a value of 0.133. We performed sensitivity analysis at the cohort level: in addition to the primary analysis based on the propensity score matching (PSM) cohort, we repeated the multivariate logistic regression analysis using the full, unmatched initial cohort, adjusting for all available confounding factors. The results were consistent with the primary analysis: bisphosphonate use remained a significant protective factor associated with a reduced risk of SSI (Table  3 ). We performed a simplified binary logistic regression analysis, which showed that medication use was an independent protective factor against postoperative infection (aOR = 0.136, 95% CI: 0.048–0.384, P  < 0.001). Simultaneously, smoking was confirmed as an independent risk factor (aOR = 3.856, 95% CI: 1.153–12.900, P  = 0.028) . Table 2 Results of binary logistic regression analysis for SSI in the propensity score-matched cohort variable OR 95%CI Pvalue Take medication 0.129 0.045—0.365 0.000*** Age 0.968 0.927—1.010 0.129 Length of hospital stay 1.025 0.978—1.074 0.309 Operative time 1.082 0.080—14.701 0.953 Intraoperative blood loss 0.978 0.884—1.081 0.661 Obesity (BMI > 28) 2.000 0.794—5.037 0.141 Hypertension 0.856 0.399—1.834 0.689 Smoking 3.856 1.123—13.231 0.032* History of diabetes mellitus 1.574 0.618—4.006 0.342 Anaemia 2.583 0.664—10.042 0.171 P value greater than 0.05, no statistical difference. The P value is less than 0.05, indicating a statistical difference with a significance of*. The P value is less than 0.001, and the significance is*** Results of binary logistic regression analysis for SSI in the propensity score-matched cohort P value greater than 0.05, no statistical difference. The P value is less than 0.05, indicating a statistical difference with a significance of*. The P value is less than 0.001, and the significance is*** Table 3 Binary logistic regression analysis of risk factors for SSI in the unmatched cohort Variable OR Exp (B) (95% C.I. of Exp (B)) Pvalue Take medication 3.498 1.252 (1.384–8.844) 0.008** Age 0.981 -0.019 (0.945–1.019) 0.317 Length of hospital stay 1. 002 0.002 (0.965–1.041) 0.904 Weight 1.013 0.012 (0.941–1.090) 0.740 Body mass index (BMI) 0.940 -0.061 (0.771–1.147) 0.544 Hypertension 0.968 -0.033 (0.500-1.871) 0.922 P value greater than 0.05, no statistical difference. P value < 0.01, significant difference** (Sig. / p   1 were risk factors and < 1 were protective factors) Binary logistic regression analysis of risk factors for SSI in the unmatched cohort P value greater than 0.05, no statistical difference. P value < 0.01, significant difference** (Sig. / p   1 were risk factors and < 1 were protective factors) We employed independent samples t-tests to compare preoperative and postoperative renal function, various peripheral blood parameters, and body temperature between the two patient groups in the propensity score-matched comparable cohort. No statistically significant differences were observed between the groups (Tables  4 and 5 , all P-values > 0.05). During the 90-day follow-up, we found no evidence that bisphosphonate use after anterior cervical discectomy and fusion (ACDF) increases related adverse reactions in postmenopausal women with postmenopausal osteoporosis. That is, there was no significant difference in adverse reactions between the medicated and non-medicated groups. Table 4 Postoperative indicators of patients variable Medication group Control group T-value Pvalue 95%CI Lower limit Upper limit Platelet count 243.88 ± 62.237 236.82 ± 67.814 -1.285 0.199 -17.858 3.729 Lymphocyte count 1.15 ± 0.484 1.20 ± 0.520 1.173 0.241 -0.033 0.133 Mononuclear cell count 0.47 ± 0.183 0.48 ± 0.191 0.787 0.432 -0.019 0.043 Temperature 36.51 ± 0.192 36.52 ± 0.192 0.248 0.804 -0.028 0.036 White blood cell count 10.58 ± 3.467 11.18 ± 3.798 1.943 0.052 -0.006 1.200 Creatinine 0.60 ± 0.061 0.60 ± 0.059 0.142 0.887 -0.009 0.010 Glomerular Filtration Rate 98.10 ± 6.478 97.80 ± 6.108 -0.589 0.556 -1.328 0.715 Blood calcium 1.12 ± 0.029 1.25 ± 0.029 1.645 0.100 -0.001 0.009 P value greater than 0.05, no statistical difference Postoperative indicators of patients P value greater than 0.05, no statistical difference Table 5 Patient preoperative indicators variable Medication group Control group T-value Pvalue 95%CI Lower limit Upper limit Platelet count 265.99 ± 65.890 262.30 ± 67.525 -0.661 0.509 -14.655 7.272 Lymphocyte count 2.093 ± 0.666 2.025 ± 0.638 -1.261 0.208 -0.174 0.038 Mononuclear cell count 0.47 ± 0.178 0.48 ± 0.174 0.427 0.669 -0.022 0.035 Temperature 36.51 ± 0.183 36.53 ± 0.223 0.939 0.348 -0.018 0.051 White blood cell count 6.50 ± 2.341 6.42 ± 1.909 -0.466 0.642 -0.419 0.259 Creatinine 0.64 ± 0.085 0.65 ± 0.082 0.877 0.381 -0.008 0.020 Glomerular Filtration Rate 94.76 ± 8.644 94.02 ± 7.956 -1.075 0.282 -2.079 0.607 Blood calcium 1.25 ± 0.030 1.25 ± 0.029 0.776 0.438 -0.007 0.003 P value greater than 0.05, no statistical difference Patient preoperative indicators P value greater than 0.05, no statistical difference

After propensity score matching to balance baseline variables, the risk of postoperative infection in medicated patients was only 13.6% of that in non-medicated patients (Table 3 ), an effect that was highly statistically significant. This indicates that the use of bisphosphonates indeed significantly reduces the infection rate after ACDF in postmenopausal women with osteoporosis. Among the confounding factors, smoking ( p < 0.01, Table 3 ) showed a statistically significant correlation with infection. This is consistent with the risk factors mentioned in the reference [ 14 ]. In this study, compared to the non-medication group, the bisphosphonate-treated group showed no statistically significant difference in the common adverse reactions associated with bisphosphonate use (including renal impairment, hypocalcemia, elevated body temperature, increased peripheral blood lymphocytes, and elevated white blood cell count). It should be emphasized that certain long-term risks associated with bisphosphonates—such as atypical femoral fractures and osteonecrosis of the jaw—were not evaluable within the follow-up period of this study. These risks are typically associated with prolonged therapy (often beyond three years), whereas the present study focused solely on early postoperative single-dose or short-term use. Therefore, our safety conclusions are applicable only to the context of short-term administration. Future studies with prolonged follow-up are recommended to comprehensively evaluate the long-term safety of bisphosphonate use. The results of this study confirm the protective effect of the medication while also identifying smoking as a risk factor for infection, which aligns with the extensive body of evidence-based medicine. This confirms, on one hand, the sensitivity of our data and model, and on the other hand, suggests that in clinical practice, advising patients to quit smoking and administering prophylactic medication may serve as synergistic strategies for reducing postoperative infection risk. In defining the exposure window for medication use, potential immortal time bias was considered. To mitigate this bias, the “medication group” was strictly defined as patients who initiated bisphosphonate therapy early after surgery (median time: postoperative day 2), with the first dose administered prior to any diagnosis of surgical site infection (SSI). This clear temporal sequence—where drug exposure consistently preceded the outcome event—strengthens the plausibility of a causal inference regarding the association between bisphosphonate use and reduced SSI risk. Nevertheless, as this is a retrospective study without randomization of treatment timing, we recommend that future prospective studies further examine the effect of medication timing on infection prevention. This study analyzed 987 patients undergoing ACDF to evaluate whether bisphosphonate use reduces postoperative infection and is safe in postmenopausal osteoporotic women. In this trial, postoperative treatment with bisphosphonates showed significant statistical proof of association with reduced postoperative infection. This supports our hypothesis. Although a higher postoperative infection rate after ACDF has been noted previously [ 15 ], to our knowledge, this is the first study to evaluate the incidence of postoperative infection after ACDF in postmenopausal patients with osteoporosis who take bisphosphonates. Although the therapeutic effect of bisphosphonates on osteoporosis is widely recognized, high-quality evidence on bisphosphonates and postoperative infection for specific surgical procedures remains very limited. This study, with a larger sample size than similar literature, a more homogeneous surgical team, and stricter control of key confounding factors, clearly confirmed the strength and persistence of this association for the first time in the special subgroup of postmenopausal women, filling a gap in evidence in this field. This provides clinicians with richer options for drug selection to prevent postoperative infection after ACDF. A major methodological strength of this study lies in the fact that all surgical procedures were performed by a highly homogeneous surgical team. This team consisted of a fixed chief spine surgeon as the primary operator, assistants, and scrub nurses, adhering to strictly uniform surgical approaches, technical details, and perioperative management protocols. This setup ensured a high degree of standardization and reproducibility of surgical techniques throughout the study period. This homogeneity minimized the ‘technical bias’ introduced by individual surgeon differences in skill, preference, or learning curve. It made the observed differences in postoperative infection more likely attributable to postoperative bisphosphonate use rather than the performer’s technical variables. Therefore, it significantly enhanced the internal validity of the study, providing a solid methodological foundation for the reliability of our core conclusion. Compared to studies involving multiple surgeons with varying degrees of technical standardization, this study provided a cleaner and more controlled technical background for evaluating the effect of postoperative bisphosphonate use on infection after Anterior Cervical Discectomy and Fusion in postmenopausal osteoporotic women. The use of anti-osteoporotic drugs is an important measure to reduce infection after fracture [ 16 ]. However, whether bisphosphonate use can reduce postoperative infection after ACDF remains unclear. This study is the first to observe a significant association between bisphosphonate use and a reduced risk of surgical site infection (SSI) following ACDF in postmenopausal osteoporotic patients. The observed reduction in infection risk may be related to the anti-inflammatory properties of bisphosphonates, and warrants further investigation in future studies. This result is consistent with the study by Eric et al., who proposed that bisphosphonates can exert antibacterial effects by inhibiting enzymes crucial for bacterial cell wall and coenzyme Q biosynthesis [ 17 ]. In vitro studies have shown that bisphosphonates can effectively inhibit the biofilm formation of common SSI pathogens—Staphylococcus aureus and Staphylococcus epidermidis [ 18 ]. Furthermore, by inhibiting the mevalonate pathway, bisphosphonates can stimulate the expansion and activation of gamma delta T (γδ T) cells [ 19 ]. While these preclinical findings suggest potential immunomodulatory and antibacterial mechanisms, our clinical study was not designed to confirm them. Consequently, the specific mechanisms by which bisphosphonates might reduce SSI risk remain hypothetical and should be explored in future translational research. Literature [ 20 , 21 ] has reported that obesity is a risk factor for postoperative complications, especially surgical site infection, so we included BMI as a control variable. Hypertensive patients may have vascular pathologies, including small vessel and microvascular issues [ 22 ]. This may lead to local ischemia and hypoxia at the incision site, ultimately causing infection or delayed healing in the ACDF site. Additionally, diabetes and anemia are important factors for postoperative infection [ 23 ]. Therefore, although the confounding factors showed no statistical difference between the two groups after SMD assessment, we still used logistic regression analysis to explore the relationship between these confounding factors, bisphosphonate use, and infection. In this study, among the confounding factors, smoking showed a clear statistical significance with postoperative infection rate, but in the logistic regression, the OR was significantly > 1. This is consistent with the conclusions of James W-MD [ 20 ] and others. We believe that orthopedic surgeons should still be vigilant about these baseline patient data. Osteoporosis patients are mostly elderly women, while young men are considered representative of normal bone density [ 24 – 26 ]. Our analysis suggests a positive correlation between age and the occurrence of postoperative infection after ACDF. This may indicate that different detection standards should be applied for the elderly and for female versus male patients [ 27 – 29 ]. Postmenopausal women are a special subgroup. Due to hormonal influences, they are not only prone to osteoporosis but also may develop endometriosis or hyperplasia due to lifestyle or diet [ 30 , 31 ]. Taking endometriosis as an example, although its active period is usually during the reproductive years, persistent or new-onset lesions after menopause highlight the special risk of abnormal tissue proliferation and uncontrolled inflammatory response in this population after drastic hormonal environmental changes. In this context, the value of any research addressing specific health issues in postmenopausal women (such as the reduction of postoperative infection risk by bisphosphonates explored in this study) is amplified. Socio-demographic factors such as low income levels and insufficient knowledge about osteoporosis can lead to discontinuation of drug therapy within 1 year of diagnosis in young osteoporosis patients [ 23 ]. Based on our data, the importance of assessing bone quality in all patients undergoing ACDF, regardless of socio-demographic factors, cannot be overemphasized. This study acknowledges several limitations inherent to its retrospective observational design. Furthermore, the single-center nature of this study and the highly standardized surgical protocol, while strengths for internal validity, may limit the generalizability of our findings to other institutions with different patient populations, surgical techniques, or perioperative care pathways. Future multi-center, prospective studies are warranted to validate our results in broader clinical settings.Selection bias exists because patients in the control group did not receive bisphosphonates due to contraindications or personal choice, reflecting real-world clinical decision-making. This makes the study conclusions directly applicable to patient groups in the real world where treatment suitability needs to be considered. Another limitation is that, due to interface issues between the hospital’s comprehensive disease platform and the electronic medical record system during certain periods, as well as limitations in the underlying data extraction logic of the platform, the medical record data for 133 patients had a missing rate exceeding 20%, leading to their exclusion from the analysis. It is important to emphasize that this data missingness was primarily due to technical reasons related to cross-system information exchange, rather than clinical factors such as patient disease severity or treatment outcomes. To assess the potential selection bias introduced by excluding these 133 patients with > 20% missing data, we compared the age variable between the included and excluded cohorts. No significant difference was found ( p  > 0.05, see Supplementary Table S1 for details). Therefore, we believe the risk of this exclusion process introducing serious selection bias is low. Although the missingness may affect the representativeness of the study cohort to the original population to some extent, considering the rigorous application of propensity score matching and the consistency of results across analyses—which supports the internal validity of the comparative results within the studied population—we believe this technical exclusion is unlikely to fundamentally alter the core conclusion of this study regarding the association between bisphosphonate use and reduced infection risk. Finally, although our sample size ( n  = 987) exceeds that of many similar studies, providing sufficient statistical power, a formal a priori sample size calculation was not performed due to the lack of reliable prior effect estimates—this is a methodological consideration for future prospective studies. Despite rigorous adjustment using PSM and multivariate regression, the possibility of residual confounding from unmeasured factors cannot be eliminated. To quantitatively assess the robustness of our findings to unmeasured confounding, we calculated the E-value for the primary association between intravenous zoledronic acid use and SSI risk. The E-value represents the minimum strength of association (on the risk ratio scale) that an unmeasured confounder would need to have with both the exposure and the outcome to fully explain away the observed association, conditional on the measured covariates. For the adjusted odds ratio of 0.136, the E-value was 14.18. This indicates that an unmeasured confounder would need to be associated with both bisphosphonate use and SSI by a risk ratio of 14.18-fold each to negate the observed protective effect, which is considerably stronger than any measured confounder in our study (e.g., smoking had an OR of 3.856). Therefore, our findings appear robust to potential unmeasured confounding.

For postmenopausal osteoporotic women undergoing ACDF, postoperative bisphosphonate use is significantly associated with a reduced risk of surgical site infection and demonstrates good short-term safety. Smoking remains a modifiable risk factor. These findings provide a new perspective for preventive strategies in this vulnerable patient population.It is important to emphasize that the favorable safety profile observed in this study pertains specifically to the 90-day postoperative period following a single dose. The long-term risks associated with bisphosphonate therapy, such as osteonecrosis of the jaw or atypical femoral fractures, were not assessed and necessitate ongoing monitoring in clinical practice.

Surgical site infection (SSI) is one of the most common postoperative complications. It can lead to prolonged hospitalization, delayed wound healing, and increased morbidity and mortality. Although surgical techniques and perioperative measures are continuously advancing and improving, adverse events in patients undergoing cervical spine surgery remain unavoidable. Surgical site infection is the third most common type of hospital-acquired infection, after urinary tract infections (23–40%) and respiratory complications (17–23%) [ 1 , 2 ].Furthermore, complex cervical spine surgical approaches (such as the transoral approach) may disrupt the mucosal barrier, exposing the surgical site to bacterial contamination and increasing the risk of infection [ 3 ]. The overall incidence of SSI after cervical spine surgery varies depending on the type of surgery and population. Common degenerative spine surgeries such as Anterior Cervical Discectomy and Fusion (ACDF) are described as a safe and efficient strategy for treating degenerative cervical spine disease [ 4 ]. It is performed in the clean space posterior to the trachea and esophagus, without disrupting the mucosal lining of the respiratory or digestive tracts, and its infection rate is generally reported to be below 2% [ 5 , 6 ]. SSI is associated with delayed wound healing, prolonged hospitalization, and increased mortality, all of which lead to significant clinical and economic burdens. Osteoporosis has been confirmed as an independent risk factor for postoperative infection [ 7 ]. Osteoporosis is a metabolic bone disease characterized by reduced bone mass and deterioration of bone microstructure. The accelerated bone resorption caused by a sharp decline in estrogen levels is its main pathological mechanism, clinically manifested as increased bone fragility and elevated fracture risk. This metabolic bone disease characterized by reduced bone mass and deterioration of bone microstructure may increase infection risk through the following mechanisms: Decreased bone quality: Destruction of trabecular structure reduces implant stability and prolongs operative time; Impaired local blood supply: Thinning of the cortical bone leads to increased intraoperative blood loss, affecting incision healing; Altered immune microenvironment: Enhanced osteoclast activity releases pro-inflammatory factors (such as IL-6, TNF-α), inhibiting neutrophil bactericidal function [ 8 ]. Currently, administering anti-osteoporotic drugs to patients during the perioperative period is widely recognized as effective in preventing bone loss during spinal fusion surgery [ 9 ].Bisphosphonates (such as alendronate, zoledronic acid), as first-line anti-osteoporotic drugs, reduce bone resorption by inhibiting osteoclast activity, significantly lowering the risk of vertebral fractures. In addition to skeletal benefits, evidence suggests they may possess anti-inflammatory properties. Their anti-inflammatory effect may be achieved by blocking the mevalonate pathway and reducing IL-1β secretion [ 10 ].Although these properties suggest they may reduce the risk of postoperative infection, clinical evidence specifically for cervical spine surgery, especially Anterior Cervical Discectomy and Fusion (ACDF), is currently lacking. Notably, the Chinese Expert Consensus on Perioperative Management of Osteoporotic Fractures (2018) mentions that cervical spine surgery carries a high risk of infection due to complex anatomy (adjacent to airway, esophagus) and postoperative dead space formation, and infection can lead to catastrophic complications (such as mediastinitis, epidural abscess). Therefore, exploring the value of bisphosphonates in infection prevention for such high-risk surgeries has important clinical significance. However, current evidence is lacking on whether bisphosphonates reduce postoperative infection after ACDF in postmenopausal women with osteoporosis. There is also a lack of evaluation of the incidence of adverse reactions to bisphosphonates after ACDF in this vulnerable population. If bisphosphonates effectively reduce postoperative infection risk without compromising safety, this finding would guide clinicians in selecting preventive medications for this vulnerable population undergoing ACDF. Therefore, it is very necessary to conduct research. This study aims to assess whether bisphosphonate use reduces the risk of SSI and evaluate its safety in postmenopausal osteoporotic patients after undergoing ACDF.

Supplementary Material 1. Supplementary Material 1.