Lumbar spinal fusion is a widely used technique in the treatment of conditions such as lumbar stenosis and degenerative spondylolisthesis. However, acute postoperative infection remains a significant complication. The aim of this study was to identify risk factors associated with a higher incidence of acute infection following posterior lumbar fusion for degenerative pathology.
Material and methodsWe conducted an observational, retrospective, and analytical study at a single hospital centre. Patients over 18 years of age who underwent instrumented lumbar fusion for lumbar stenosis or degenerative spondylolisthesis between 2017 and 2022 were included. Demographic, clinical, and surgical variables were analysed. A bivariate analysis was followed by a multivariate logistic regression model to identify independent risk factors for acute postoperative infection.
ResultsA total of 356 patients were analysed, with a mean age of 60.4years; 55.6% were male. The overall incidence of acute postoperative infection was 5.9% (21cases). Bivariate analysis revealed significant associations with infection for diabetes mellitus (P=.010), hypertension (P=.042), chronic corticosteroid use (P=.001), and surgical duration>3h (P=.028). In the multivariate model, only diabetes mellitus remained a significant independent risk factor (OR: 3.36; 95% CI: 1.22–9.31; P=.019).
ConclusionsDiabetes mellitus was identified as the main independent risk factor for acute postoperative infection. This finding highlights the importance of a thorough preoperative assessment of surgical risk in diabetic patients and the need to provide them with adequate information during the informed consent process.
La artrodesis lumbar es una técnica ampliamente utilizada en el tratamiento de patologías como la estenosis lumbar y la espondilolistesis degenerativa. Sin embargo, la infección aguda posquirúrgica sigue siendo una complicación relevante. El objetivo de este estudio fue identificar los factores de riesgo asociados con una mayor incidencia de infección aguda tras una artrodesis lumbar posterior por patología degenerativa.
Material y métodosSe realizó un estudio observacional, retrospectivo y analítico en un único centro hospitalario. Se incluyeron pacientes mayores de 18años sometidos a artrodesis lumbar instrumentada por estenosis o espondilolistesis degenerativa entre 2017 y 2022. Se analizaron variables demográficas, clínicas y quirúrgicas. Se realizó un análisis bivariado seguido de un modelo multivariante de regresión logística para identificar factores de riesgo independientes de infección aguda postoperatoria.
ResultadosSe analizaron 356 pacientes, con una edad media de 60,4años; el 55,6% eran hombres. La incidencia global de infección aguda posquirúrgica fue del 5,9% (21casos). En el análisis bivariado, se asociaron significativamente con infección: diabetes mellitus (p=0,010), hipertensión arterial (p=0,042), uso crónico de corticoides (p=0,001) y duración quirúrgica >3horas (p=0,028). En el análisis multivariante, solo la diabetes mellitus se mantuvo como factor de riesgo independiente (OR: 3,36; IC95%: 1,22-9,31; p=0,019).
ConclusionesLa diabetes mellitus se identificó como el principal factor de riesgo independiente para infección aguda posquirúrgica. Este hallazgo resalta la importancia de una evaluación preoperatoria exhaustiva del riesgo quirúrgico en pacientes diabéticos y su adecuada información en el proceso de consentimiento informado.
Lumbar arthrodesis with pedicle screws and/or interbody cages is a widely used surgical procedure for treating degenerative spinal conditions, such as lumbar spinal stenosis or degenerative spondylolisthesis, with generally satisfactory results in terms of pain relief and functional recovery.1,2 However, postoperative infectious complications continue to represent one of the most significant challenges in this area, with an incidence ranging from 1% to 13% according to reported series.3–5 These infections, especially when they occur acutely, are associated with a significant increase in morbidity, the need for reoperations, prolonged hospital stays, and high healthcare costs.6,7 Therefore, identifying modifiable or controllable risk factors becomes a priority to improve outcomes and reduce the clinical and economic impact of these complications.5,8,9 The risk factors most consistently reported in the literature include: diabetes mellitus; obesity; smoking; prolonged surgery duration; the number of instrumented levels, and corticosteroid use.4,5,8,10–12
In particular, diabetes has been identified as a relevant factor in multiple surgical procedures, and its relationship with infection appears to be mediated by micro vascular alterations, immune function deficits, and impaired wound healing.13–15 Despite the abundant literature on risk factors for infection in spinal surgery, specific evidence on the role of diabetes as an independent predictor in instrumented lumbar arthrodesis for degenerative disease is still limited and heterogeneous: many studies include mixed populations with different pathologies and types of surgery. Therefore, the main objective of this study was to determine the risk factors associated with a higher incidence of acute infection after instrumented lumbar arthrodesis in patients operated on for degenerative lumbar disease.
Material and methodsStudy designA retrospective, analytical, observational study was conducted at a single hospital, including patients who underwent surgery consecutively between January 2017 and December 2022.
Inclusion and exclusion criteriaPatients over 18 years of age who underwent instrumented lumbar surgery for degenerative conditions (lumbar stenosis and degenerative spondylolisthesis) were included. Patients with iliac instrumentation, instrumentation involving more than 5 levels, scoliosis, adjacent segment syndrome (previous spinal fusion extensions), anterior lumbar arthrodesis, and incomplete medical records were excluded
Study variablesDemographic variables included age, sex, and body mass index (BMI). Clinical variables considered were history of previous surgery (including microdiscectomy, foraminotomy, and previous spine surgery without instrumentation), diabetes mellitus, hypertension, chronic corticosteroid use, smoking, and American Society of Anaesthesiologists (ASA) classification. Surgical variables analysed included the reason for surgery (lumbar stenosis versus degenerative spondylolisthesis); type of surgery (posterior arthrodesis with foraminotomy; posterior arthrodesis with laminectomy; circumferential arthrodesis with foraminotomy; circumferential arthrodesis with laminectomy); number of levels operated; duration of surgery; postoperative transfusion, and use of postoperative redon drains. The primary outcome of the study was the development of acute deep postoperative infection, defined as infection occurring within the first 8 weeks after surgery. Initial diagnosis included: the presence of purulent drainage at the incision site or in deep spaces; formation of fistulous tracts; demonstration of abscess collections by imaging studies or discovery during surgical reintervention; identification of pathogens in two or more culture samples obtained during the surgical procedure; positive sonication with ≥50CFU/ml, and a clinical diagnosis established by the attending surgeon based on preoperative and intraoperative signs and symptoms (hyperthermia, erythema, evidence of deep tissue involvement, and/or purulent discharge.16,17
Statistical analysisA descriptive statistical analysis was performed on all variables. Categorical variables were expressed as absolute and relative frequencies (percentages), while continuous variables were expressed as mean±standard deviation. For bivariate analysis, the Chi-square test was used for categorical variables and the Student's t-test for continuous variables. A p-value<.05 was considered statistically significant.
Subsequently, a multivariate analysis was performed using logistic regression, including all variables that showed statistical significance in the bivariate analysis (p<.05). The results are expressed as odds ratios (OR) with 95% confidence intervals (95% CI). Statistical analysis was performed using SPSS version 25.0 (IBM Corp., Armonk, New York, USA).
ResultsA total of 356 patients were included in the study, of whom 198 (55.6%) were men (Table 1). The mean age was 60.4 years (SD±12.4). 73.0% underwent surgery for lumbar stenosis and 27.0% for degenerative spondylolisthesis. Regarding the surgical technique, the most frequent procedure was posterior arthrodesis with laminectomy (48.9%), followed by circumferential arthrodesis with foraminotomy (28.1%).
Description of the general characteristics of the study participants.
| Acute postoperative infection | ||||
|---|---|---|---|---|
| No (n=335) | Yes (n=21) | Total (n=356) | p | |
| Sex | .550 | |||
| Man, % (n) | 55.2 (185) | 61.9 (13) | 55.6 (198) | |
| Woman, % (n) | 44.8 (150) | 38.1 (8) | 44.4 (158) | |
| Age (years) | .301 | |||
| Mean (SD) | 60.3 (12.3) | 63.1 (13.3) | 60.4 (12.4) | |
| BMI | .495 | |||
| Mean (SD) | 28.2 (4.5) | 28.9 (4.9) | 28.2 (4.6) | |
| Previous surgery | .308 | |||
| No, % (n) | 80.6 (270) | 71.4 (15) | 80.1 (285) | |
| Yes, % (n) | 19.4 (65) | 28.6 (6) | 19.9 (71) | |
| Reason for surgery | .737 | |||
| Lumbar stenosis. % (n) | 72.8 (244) | 76.2 16) | 73.0 (260) | |
| Degenerative spondylolisthesis. % (n) | 27.2 (91) | 23.8 (5) | 27.0 (96) | |
| Type of surgery | .426 | |||
| Posterior arthrodesis+Foraminotomy. % (n) | 6.3 (21) | 4.8 (1) | 6.2 (22) | |
| Posterior arthrodesis+Laminectomy. % (n) | 48.1 (161) | 61.9 (13) | 48.9 (174) | |
| Circumferential arthrodesis+Foraminotomy. % (n) | 28.1 (94) | 28.5 (6) | 28.1 (100) | |
| Circumferential arthrodesis+Laminectomy. % (n) | 17.5 (59) | 4.8 (1) | 16.8 (60) | |
| Number of levels operated | .194 | |||
| 1. % (n) | 48.5 (162) | 52.4 (11) | 48.6 (173) | |
| 2. % (n) | 37.4 (125) | 19.0 (4) | 36.2 (129) | |
| 3. % (n) | 11.4 (38) | 23.8 (5) | 12.2 (43) | |
| 4. % (n) | 2.7 (9) | 4.8 (1) | 2.8 (10) | |
| BMI≥30 | .078 | |||
| No, % (n) | 66.5 (220) | 47.6 (10) | 65.3 (230) | |
| Yes, % (n) | 33.5 (111) | 52.4 (11) | 34.7 (122) | |
| ASA scale≥III | .133 | |||
| No, % (n) | 80.3 (269) | 66.7 (14) | 79.5 (283) | |
| Yes, % (n) | 19.7 (66) | 33.3 (7) | 20.5 (73) | |
| Diabetes mellitus | .010 | |||
| No, % (n) | 83.9 (281) | 61.9 (13) | 82.6 (294) | |
| Yes, % (n) | 16.1 (54) | 38.1 (8) | 17.4 (62) | |
| High blood pressure | .042 | |||
| No, % (n) | 60.6 (203) | 38.1 (8) | 59.3 (211) | |
| Yes, % (n) | 39.4 (132) | 61.9 (13) | 40.7 (145) | |
| Use of corticoids | .001 | |||
| No, % (n) | 97.9 (328) | 85.7 (18) | 97.2 (346) | |
| Yes, % (n) | 1.1 (7) | 14.3 (3) | 2.8 (10) | |
| Smoker | .373 | |||
| No, % (n) | 71.0 (238) | 61.9 (13) | 70.5 (251) | |
| Yes, % (n) | 29.0 (97) | 38.1 (8) | 29.5 (105) | |
| Duration of IQ>3h | 0.028 | |||
| No, % (n) | 69.6 (156) | 44.4 (8) | 67.8 (164) | |
| Yes, % (n) | 30.4 (68) | 55.6 (10) | 32.2 (78) | |
| Postoperative transfusion | .629 | |||
| No, % (n) | 89.1 (295) | 85.7 (18) | 88.9 (313) | |
| Yes, % (n) | 10.9 (36) | 14.3 (3) | 11.1 (39) | |
| Use of redon drain | .825 | |||
| No, % (n) | 91.8 (304) | 90.5 (19) | 91.8 (323) | |
| Yes, % (n) | 8.2 (27) | 9.5 (2) | 8.2 (29) | |
ASA: American Society of Anaesthesiologists; BMI: body mass index; IQ: surgery; n: number of patients; %: percentage.
The overall incidence of deep postoperative acute infection was 5.9% (21 cases). All patients diagnosed with deep postoperative acute infection in our sample were treated with surgical debridement and initially empirical antibiotic therapy, which was subsequently targeted based on antibiogram results. The most frequently isolated microorganism was Staphylococcus epidermidis (38%), followed by Staphylococcus aureus (34%) and polymicrobial infections (34%). The mean intervention time was 147.3±53.9min, with an interquartile range (IQR) of 70min.
In the bivariate analysis, the following factors were identified as significantly associated with a higher risk of infection: diabetes mellitus (p=.010), hypertension (p=.042), chronic corticosteroid use (p=.001), and a surgical duration greater than 3hours (p=.028). A BMI≥30 showed no significant trend (p=.078) (Table 2).
Risk factors associated with postoperative infection in posterior lumbar arthrodesis for degenerative pathology. Bivariate analysis.
| Bivariate | ||
|---|---|---|
| OR (95% CI) | p | |
| Age | 1.02 (.98–1.06) | .301 |
| Sex | ||
| Man | 1 | |
| Woman | .76 (.31–1.88) | .551 |
| BMI | 1.03 (.94–1.14) | .494 |
| Previous surgery | ||
| No | 1 | |
| Yes | 1.66 (.62–4.45) | .312 |
| Reason for surgery | ||
| Degenerative spondylolisthesis | 1 | |
| Lumbar stenosis | 1.19 (.43–3.35) | .737 |
| Type of surgery | ||
| Posterior arthrodesis+Foraminotomy | 1 | |
| Posterior arthrodesis+Laminectomy | 1.69 (.21–13.63) | .619 |
| Circumferential arthrodesis+Foraminotomy | 1.34 (.15–11.73) | .791 |
| Circumferential arthrodesis+Laminectomy | .36 (.21–5.95) | .472 |
| Number of operational levels | ||
| 1 | 1 | |
| 2 | .47 (.15–1.52) | .207 |
| 3 | 1.94 (.64–5.91) | .245 |
| 4 | 1.64 (.19–14.11) | .654 |
| BMI≥30 | ||
| No | 1 | |
| Yes | 2.18 (.89–5.29) | .085 |
| ASA Scale≥III | ||
| No | 1 | |
| Yes | 2.04 (.79–5.25) | .140 |
| Diabetes mellitus | ||
| No | 1 | |
| Yes | 3.20 (1.27–8.10) | .014 |
| High blood pressure | ||
| No | 1 | |
| Yes | 2.50 (1.01–6.19) | .048 |
| Use of corticoids | ||
| No | 1 | |
| Yes | 7.81 (1.86–32.74) | .005 |
| Smoker | ||
| No | 1 | |
| Yes | 1.51 (.61–3.76) | .376 |
| Duration of IQ>3h | ||
| No | 1 | |
| Yes | 2.87 (1.09–7.58) | .034 |
| Postoperative transfusion | ||
| No | 1 | |
| Yes | 1.37 (.38–4.87) | .631 |
| Use of redon drain | ||
| No | 1 | |
| Yes | 1.18 (.26–5.36) | .825 |
ASA: American Society of Anesthesiologists; BMI: body mass index; CI: confidence interval; IQ: surgery; n: number of patients; OR: odds ratio; %: percentage.
In the multivariate logistic regression model, only diabetes mellitus remained an independent and significant risk factor for acute postoperative infection (OR: 3.36; 95% CI: 1.22–9.31; p=.019). Although the use of corticosteroids (OR: 2.95; p=.233) and surgery durations longer than 3h (OR: 2.37; p=.092) showed trends, they did not reach statistical significance in this analysis (Table 3).
Risk factors associated with postoperative infection in posterior lumbar arthrodesis for degenerative pathology. Multivariate analysis.
| Multivariate | ||
|---|---|---|
| OR (95% CI | p | |
| Diabetes mellitus | ||
| No | 1 | |
| Yes | 3.36 (1.22–9.31) | .019 |
| Use of corticoids | ||
| No | 1 | |
| Yes | 2.95 (.50–17.48) | .233 |
| Duration of IQ>3h | ||
| No | 1 | |
| Yes | 2.37 (.87–6.48) | .092 |
CI: confidence interval; IQ: surgery; OR: odds ratio.
Lumbar arthrodesis with pedicle screws and/or interbody cages has experienced exponential growth in recent decades, becoming the treatment of choice for multiple degenerative conditions, such as lumbar stenosis and degenerative spondylolisthesis.1,2,18,19 However, this type of surgery is not without complications, with acute postoperative infection being one of the most feared due to its impact on morbidity, mortality, and healthcare costs.3–7 The main finding of our study was that diabetes mellitus was identified as the only independent risk factor in the multivariate analysis, with an odds ratio (OR) of 3.36, implying that diabetic patients have an approximately three-fold increased risk of developing an acute postoperative infection.
The infection rate of 5.9% observed in our study is comparable to that reported by other authors in similar series. Blumberg et al. reported an incidence of 4.7% in a series of 3101 patients undergoing spinal fusion,7 while Nota et al. found a rate of 6.1% in 5791 spinal fusions.20 Rico Nieto et al. diagnosed 41 cases of infection associated with spinal instrumentation out of a total of 1680 procedures, with a mean annual incidence of 2.43%.21 Kurtz et al., in a 10-year follow-up, reported a cumulative incidence of 8.5% in primary lumbar arthrodesis.22 Yaldiz et al. found a rate of 11.6% in a series of 540 patients who underwent lumbar arthrodesis.23 Chaichana et al. found an infection rate of 4.5% after analysing 817 lumbar fusions for degenerative pathology.10 Regarding microbiology, the predominance of Staphylococcus epidermidis (38%) and Staphylococcus aureus (34%) in our study is consistent with previous literature. Several studies have identified staphylococci as the most frequent pathogens in surgical site infections in spinal surgery, representing between 60–80% of cases.3,4,9,10,17,21,24 This consistency in infection rates and microbiology suggests that our results are representative of current clinical practice.
The most relevant finding of our study was the identification of diabetes mellitus as the only independent risk factor for acute postoperative infection following instrumented lumbar surgery, with an odds ratio (OR) of 3.36 (95% CI: 1.22–9.31; p=.019) in the multivariate analysis. This result is consistent with multiple previous studies that have demonstrated the role of diabetes as a major risk factor for infection in spinal surgery.5,8,11,12,21,25 Koutsoumbelis et al., in their multivariate analysis of 3218 patients, found that diabetes increased the risk of infection in spinal surgery, with an OR of 3.20 (95% CI: 1.22–8.40; p=.018).11 Olsen et al. reported an OR of 3.5 (95% CI: 1.2–10.0; p=.020) for postoperative infection in diabetic patients undergoing spinal surgery in their multivariate analysis.25 Similarly, Liu et al.12 and Chaichana et al.,10 in their respective multivariate analyses of cohorts of 2715 and 817 patients undergoing posterior lumbar surgery, found that diabetes mellitus was significantly associated with a higher risk of postoperative infection. Liu et al. reported an OR of 2.23 (95% CI: 1.1–4.5; p=0.026), while Chaichana et al. identified an OR of 5.58 (95% CI: 1.3–19.7; p=.02), thus confirming the role of this comorbidity as a relevant risk factor.10,12 These results have important clinical implications, as they reinforce the need for a thorough preoperative assessment of surgical risk in patients with diabetes mellitus. These patients should receive specific and detailed counselling about the increased risk of infection after lumbar arthrodesis in the context of the informed consent process.26 Another important point is that raised by Luo et al. in their meta-analysis on the impact of diabetes on spinal surgery, where they note that, despite being a clinically relevant issue, in practice some surgeons may not be fully aware of the link between diabetes and postoperative infection after spinal arthrodesis.15 Furthermore, several studies have shown that many professionals do not pay sufficient attention to whether preoperative blood glucose levels are within appropriate ranges or to the stability of perioperative glycaemic control.12,15,25 Therefore, the findings of our study highlight the importance of continuing to promote awareness and education regarding the impact of diabetes after lumbar arthrodesis.
Study limitationsOur study has several limitations that should be considered when interpreting the results. First, the retrospective design prevents us from confirming a direct causal relationship, although it does allow us to identify strong and clinically relevant associations, such as the one observed between diabetes mellitus and postoperative infection. Second, the relatively small sample size and the low number of events (21 infections) may have limited the statistical power to detect associations with less prevalent risk factors. This could explain why, unlike previous studies, in our multivariate analysis, prolonged surgery duration did not reach statistical significance as an independent factor.12 In our series, surgical time was relatively homogeneous, with a mean of 147.3min. However, the odds ratio (OR) of 2.37 for surgical procedures longer than 3h suggests a clinically relevant trend that could be confirmed in studies with larger sample sizes. Finally, it was not possible to discriminate between different types of diabetes (type 1 vs. type 2), treatment regimens (insulin vs. non-insulin), or levels of glycaemic control, which could have provided greater depth to the analysis. Previous studies have shown that perioperative glycaemic variability is closely associated with complications such as poor wound healing, as reported by Chen et al.27 Similarly, Peng et al. found that preoperative glycated haemoglobin (HbA1c), postoperative fasting glucose, and postoperative postprandial glucose were risk factors for infection after lumbar surgery in diabetic patients.28 Although the lack of data on glycaemic control may be considered a significant limitation, it is important to emphasise that diabetes mellitus, as a disease in itself, is already a significant risk factor for infection, regardless of the degree of control. Guzmán et al. demonstrated that both controlled and uncontrolled diabetic patients had a higher risk of postoperative infection in lumbar degenerative surgery, with an effect that was amplified in those with poor glycaemic control (OR=1.36 in controlled diabetics vs. OR=2.61 in uncontrolled diabetics).14 Therefore, current evidence supports the need to optimise preoperative and perioperative glycaemic control. In the preoperative blood tests at our centre, during the study period, only preoperative fasting blood glucose was included, without routine determination of glycosylated haemoglobin, nor routine postoperative controls. These data were therefore not collected systematically. Although this absence limits the level of precision of the analysis, we believe it does not invalidate the main finding, given that the available evidence consistently demonstrates that diabetes mellitus, regardless of its type and control, is associated with a significant increase in the risk of postoperative infection, wound complications, and adverse outcomes after spinal fusion surgery.29–31 However, we recognise that the inclusion of these parameters would allow for a better risk definition and propose their systematic incorporation in future prospective studies. Currently, it is recommended to maintain an HbA1c<7.5% and perioperative glucose<150mg/dL, with strict monitoring and rapid intervention strategies to prevent hyperglycaemic episodes throughout the perioperative period.14,15,32,33 We believe that future prospective studies should include variables related to glycaemic control, such as HbA1c, perioperative glycaemia, and glycaemic variability, in order to assess their direct impact on infection rates and establish more specific management strategies in diabetic patients undergoing spinal surgery.
ConclusionDiabetes mellitus was identified as the only independent risk factor for acute postoperative infection following instrumented lumbar arthrodesis for degenerative disease, with a three-fold increased risk. These findings reinforce the importance of a thorough preoperative assessment of surgical risk, as well as clear and specific information for diabetic patients in the context of informed consent, especially regarding the increased risk of associated infection.
Level of evidenceLevel of evidence iii.
Ethical responsibilitiesWe declare that the authors have taken into account the “Ethical Responsibilities,” including: (a) that the procedures followed in the research were in accordance with the ethical standards of the responsible human or animal experimentation committee (institutional or regional) and with the World Medical Association and the Declaration of Helsinki; (b) that the right of their patients to privacy and confidentiality was guaranteed in accordance with the relevant section of these rules, and that any type of identifying data in text or images was avoided in the article and, in any case, (c) that they are in possession of the informed consent of the patients for participation in the study and the publication of the results in printed and electronic (internet) format in the Spanish Journal of Orthopaedic Surgery and Traumatology.
FundingWe did not receive any grant funding for this article.
Conflict of interestsThe authors have no conflict of interests to declare.
