Risk factors for surgical site infection after open transforaminal lumbar interbody fusion in treating degenerative lumbar diseases

Abstract

Open transforaminal lumbar interbody fusion (O-TLIF) is an acknowledged surgical technique for addressing lumbar degenerative diseases. Postoperative surgical site infections (SSIs) can occur. This study aims to identify the independent risk factors for SSIs following O-TLIF in the treatment of lumbar spinal stenosis. A cohort of 1000 patients who underwent O-TLIF for lumbar spinal stenosis between 2013 and 2019 in our hospital was studied. These patients were divided into 2 groups: the infection group (n = 23) and the non-infection group (n = 977). Demographic, surgical, and radiological data were prospectively recorded to investigate the independent risk factors for SSI. Quantitative data were expressed as x ± s and analyzed using the independent sample t test; categorical data were presented as frequency (percentage) and compared using the chi-square test. Both univariate and multivariate logistic regression analyses were utilized to evaluate the significance of potential risk factors, and line charts and receiver operating characteristic curves were generated. The SSI occurrence rate in patients undergoing O-TLIF was 2.3% (23/1000). Multifactorial regression analysis identified older age (odds ratio [OR], 1.073; 95% confidence interval [CI], 1.020–1.129; P = .006), longer surgical duration (OR, 1.008; 95% CI, 1.002–1.013; P = .009), greater intraoperative blood loss (OR, 1.001; 95% CI, 1.000–1.002; P = .030), smoking (OR, 3.149; 95% CI, 1.278–7.758; P = .013), and diabetes (OR, 3.914; 95% CI, 1.513–10.127; P = .005) as independent risk factors for postoperative infection. The area under the curve values for age, surgical duration, intraoperative blood loss, diabetes, and smoking in predicting postoperative infection were 0.754 (0.658–0.849) (P < .001), 0.743 (0.627–0.860) (P < .001), 0.692 (0.569–0.814) (P = .002), 0.654 (0.528–0.780) (P = .012), and 0.622 (0.503–0.741) (P = .045), respectively. The optimal cutoff values were 67.5 years for age, 156 minutes for surgical duration, and 475 mL for intraoperative blood loss. Advanced age, smoking, prolonged surgical duration, diabetes, and increased intraoperative blood loss serve as independent risk factors for SSI following O-TLIF. Patients above 68 years old, with surgical durations exceeding 156 minutes, intraoperative blood losses surpassing 475 mL, those with diabetes, and smokers are at significant risk for SSI after lumbar surgery.

1. Introduction

With the aging population, the incidence of lumbar degenerative changes is on the rise, leading to back pain, leg pain, numbness, and in severe cases, cauda equina syndrome. These issues are among the most common reasons for surgical treatment in individuals over the age of 65. Open transforaminal lumbar interbody fusion (O-TLIF) has become one of the most commonly performed procedures for lumbar intervertebral fusion due to its easier anatomical exposure, removal of intervertebral discs via a foraminal approach, and minimal need for excessive traction on nerve roots and the dural sac.^[1]

Surgical site infection (SSI) post-spine surgery, especially after O-TLIF, is a significant complication. It can lead to chronic back pain, failure of internal fixation, pseudarthrosis, permanent neurological impairment, and other infectious complications, including urinary tract infections, pneumonia, and sepsis. These complications increase the duration of hospital stays and medical costs.^[2] The Centers for Disease Control and Prevention in the United States define SSI as an infection at the incision site,^[3,4] which can be categorized as superficial incisional infection, deep incisional infection, or organ/space infection. Superficial incisional infections involve the skin and subcutaneous tissue and are characterized by purulent drainage or positive sterile culture. Deep incisional infections involve muscle and fascia and are characterized by purulent drainage from deep tissue or a positive microbial culture. Organ/space infections occur deeper than the muscle and fascia and follow the same criteria as deep infections.

Past research indicates that the incidence of SSI following spinal O-TLIF ranges from 2% to 5%.^[5–8] Other studies have reported the risk of SSI post-TLIF to be as high as 8.5%.^[9,10] Therefore, effectively identifying SSI risk factors, early detection and prediction in high-risk postoperative patients, controlling modifiable risk factors, and focused monitoring of high-risk patients are vital for improving prognosis and preventing infection.

2. Materials and methods

2.1. Data collection

This research was conducted at the General Hospital of the Northern Theater Command and had obtained approval from the Institutional Review Board (IRB number: Y-2019-01). It was a retrospective observational study, with all data being anonymously collected and analyzed. Data were gathered and organized for 1047 patients who underwent O-TLIF surgery at our hospital from January 2013 to December 2019. The inclusion criteria encompassed: patients with degenerative lumbar diseases, including lumbar disc herniation, lumbar spinal stenosis, and lumbar spondylolisthesis; there are nerve root symptoms, consistent with imaging, which have failed with conservative treatment; patients with significant disc collapse, presence of instability in the power position or significant narrowing of the neural foramina; those who underwent open lumbar posterior exploration, nucleotomy, interbody fusion with bone grafting, and pedicle screw internal fixation. The exclusion criteria were as follows: incomplete clinical data; preexisting infection in other organs before the surgery; abnormal preoperative counts of white blood cells, erythrocyte sedimentation rates (ESR), and C-reactive protein levels (CRP); patients requiring simultaneous surgeries for other coexisting conditions or concurrent fractures. Following these inclusion and exclusion criteria, a total of 1000 patients were enrolled in the study, as depicted in (Fig. 1). SSI was defined as the presence of purulent drainage from the incision or deep tissues, or a positive microbiological culture within 30 days following the O-TLIF procedure.

Figure 1.

Flowchat of the study population. TLIF = transforaminal lumbar interbody fusion, SSI = surgical site infection.

2.2. Assessment factors

Preoperative general data assessment included gender, age, height, weight, body mass index, smoking status, alcohol consumption, history of previous lumbar surgery, history of high-intensity physical labor, preoperative Visual Analog Scale, liver and kidney functions, laboratory tests, and major medical comorbidities like diabetes, hypertension, heart disease, etc. Surgical and preoperative laboratory data assessment included surgical duration, number of lumbar surgery segments, time of day of surgery (morning [08:00–11:00], noon [11:00–13:00], afternoon [13:00–17:00], and evening [after 17:00]), intraoperative blood loss, ESR, preoperative CRP, preoperative white blood cells, hemoglobin, hematocrit, platelet crit, international normalized ratio, prothrombin time, activated partial thromboplastin time, thrombin time, and fibrinogen; preoperative systolic and diastolic pressures.

2.3. Method

Data analysis was performed using IBM SPSS software (version 26.0, SPSS Inc., Chicago). Quantitative data following a normal distribution were expressed as mean ± standard deviation (x ± s), and comparisons between 2 groups were made using independent samples t tests. Categorical data were presented as frequencies (percentages) and compared using the chi-square test. We considered risk factors for SSI previously identified in studies, along with factors we believed could lead to SSI, such as certain preoperative laboratory indices, surgical factors (duration and timing of surgery), history of high-intensity physical labor, and previous lumbar surgery. Univariate and multivariate logistic regression analyses were used, followed by line plot creation. For receiver operating characteristic (ROC) curve analysis, outcome variables were set as the presence or absence of postoperative infection, and test variables were indicators showing differences in multivariate regression. A P value of <.05 was considered statistically significant.

3. Results

3.1. Comparison of general patient data between groups

The incidence of surgical site infection (SSI) in patients undergoing O-TLIF surgery was 2.3% (23/1000). Patients were categorized into 2 groups: the SSI group (n = 23) and the non-SSI group (n = 977). Using independent sample t tests and chi-square tests, no statistically significant differences were observed between the 2 groups in terms of gender, height, weight, body mass index, alcohol history, history of hypertension, coronary heart disease, history of high-intensity physical labor, and previous lumbar surgery (P > .05). The non-infected group had a lower average age, smoking history, and diabetes prevalence compared to the infected group (P < .05) (Table 1).

Table 1.

Characteristics of the study population.

Demographics	Total (n = 1000)	No SSI (n = 977)	SSI (n = 23)	χ2/t value	P value
Gender, n (%)				0.114	.736
Male	513 (51.3)	502(51.4)	11 (47.8)
Female	487 (48.7)	475 (48.6)	12 (52.2)
Age	53.99 ± 14.78	53.72 ± 14.76	65.70 ± 10.53	-5.335	<.001
Height	167.06 ± 8.85	167.09 ± 8.88	165.87 ± 7.34	0.652	.514
Weight	70.20 ± 13.14	70.20 ± 13.13	70.35 ± 13.88	-0.053	.958
Preoperative VAS	4.57 ± 1.07	4.57 ± 1.08	4.62 ± 1.04	-0.207	.836
BMI	25.09 ± 3.83	25.08 ± 3.83	25.40 ± 3.75	-0.399	.690
Smoking, n (%)				6.131	.013
Yes	326 (32.6)	313 (32.0)	13 (56.5)
No	674 (67.4)	664 (68.0)	10 (43.5)
Drinking, n (%)				0.218	.640
Yes	262 (26.2)	255 (26.1)	7 (30.4)
No	738 (73.8)	722 (73.9)	16 (69.6)
Diabetes, n (%)				12.481	<.001
Yes	178 (17.8)	167 (17.1)	11 (47.8)
No	822 (82.2)	810 (82.9)	12 (52.2)
Hypertension, n (%)				0.299	.585
Yes	338 (33.8)	329 (33.7)	9 (39.1)
No	662 (66.2)	648 (66.3)	14 (60.9)
CHD, n (%)				0.133	.715
Yes	127 (12.7)	123 (12.6)	4 (17.4)
No	872 (87.3)	853 (87.4)	19 (82.6)
History of high-intensity labor, n (%)				0.064	.800
Yes	583 (58.3)	569 (58.2)	14 (60.9)
No	417 (41.7)	408 (41.8)	9 (39.1)
History of lumbar surgery, n (%)				0.782	.376
Yes	64 (6.4)	61 (6.3)	3 (13.0)
No	935 (93.6)	915 (93.7)	20 (87.0)
Liver and kidney function, n (%)				0.051	.821
Mildly	95 (9.5)	92 (9.4)	3 (13.0)
Normal	905 (90.5)	885 (90.6)	20 (87.0)

BMI = body mass index, CHD = coronary heart disease, SSI = surgical site infection, VAS = Visual Analogue Scale.

3.2. Comparison of surgical and clinical data between groups

Based on independent sample t tests and chi-square tests, both groups were compared for preoperative Visual Analog Scale scores, surgery duration, number of surgical segments, surgery date, intraoperative blood loss, liver and kidney function, ESR, preoperative CRP, preoperative white blood cell count, preoperative hemoglobin, hematocrit, platelet crit, international normalized ratio, prothrombin time, activated partial thromboplastin time, thrombin time, fibrinogen, preoperative systolic, and diastolic blood pressure. No significant differences were found (P > .05). The non-infected group had shorter surgery durations, lower intraoperative blood loss, and lower ESR compared to the infected group (P < .05) (Table 2).

Table 2.

Comparison of surgical clinical data.

Norms	Total (n = 1000)	No SSI (n = 977)	SSI (n = 23)	χ2/t value	P value
Operation time	127.11 ± 60.17	125.33 ± 57.74	202.48 ± 102.26	-3.605	.002
Number of operated vertebrae, n (%)				4.166	.384
1	545 (54.5)	536 (54.9)	9 (39.1)
2	304 (30.4)	297 (30.4)	7 (30.4)
3	114 (11.4)	109 (11.2)	5 (21.7)
4	23 (2.3)	22 (2.3)	1 (4.3)
5	14 (1.4)	13 (1.3)	1 (4.3)
Duration of operation, n (%)				1.851	.604
Morning	226 (22.6)	221 (22.6)	5 (21.7)
Noon	124 (12.4)	119 (12.2)	5 (21.7)
Afternoon	575 (57.5)	563 (57.6)	12 (52.2)
Evening	75 (7.5)	74 (7.6)	1 (4.3)
Intraoperative blood loss	336.81 ± 143.49	367.95 ± 143.17	713.04 ± 256.82	-2.435	.023
ESR	10.26 ± 10.65	10.16 ± 10.52	14.61 ± 14.78	-1.984	.048
Preoperative CRP	3.38 ± 7.24	3.30 ± 7.03	6.68 ± 12.13	-1.231	.231
Preoperative WBC	6.37 ± 1.77	6.37 ± 1.77	6.43 ± 2.04	-0.142	.887
Preoperative HB	137.52 ± 16.80	137.61 ± 16.83	133.61 ± 15.19	1.129	.259
HCT	0.42 ± 0.04	0.42 ± 0.04	0.40 ± 0.04	1.404	.161
PCT	0.23 ± 0.05	0.23 ± 0.05	0.23 ± 0.07	-0.019	.985
INR	0.97 ± 0.07	0.97 ± 0.07	0.98 ± 0.05	-0.866	.387
PT	12.22 ± 1.09	12.22 ± 1.09	12.20 ± 1.06	0.102	.919
APTT	31.10 ± 5.21	31.10 ± 5.15	31.13 ± 7.47	-0.035	.972
TT	17.21 ± 1.13	17.21 ± 1.12	17.13 ± 1.63	0.303	.762
FIB	2.99 ± 0.78	2.98 ± 0.77	3.23 ± 1.18	-1.477	.140
PSBP	133.85 ± 14.48	133.80 ± 14.52	135.55 ± 13.03	-0.557	.578
PDBP	79.67 ± 10.44	79.66 ± 10.29	80.14 ± 15.54	-0.211	.833

APTT = activated partial thromboplasting time, CRP = C-reactive protein, ESR = erythrocyte sedimentation rate, FIB = fibrinogen, HB = hemoglobin, HCT = hematocrit, INR = international normalized ratio, PCT = plateletcrit, PDBP = preoperative diastolic blood pressure, PSBP = preoperative systolic blood pressure, PT = prothrombin time, SSI = surgical site infection, TT = thrombin time, WBC = white blood cell.

3.3. Univariate and multivariate regression analysis for postoperative infection outcome

The status of postoperative infection (assigned as yes = 1, no = 0) was used as a dependent variable. Age, duration of surgery, intraoperative blood loss, smoking status (yes = 1, no = 0), and diabetes status (yes = 1, no = 0) were used as independent variables in binary logistic regression analysis, with the “input” method selected for variable entry. The univariate regression analysis identified older age, longer surgery duration, increased intraoperative blood loss, higher preoperative CRP, smoking, and diabetes as significant risk factors for postoperative infection. Multivariate regression analysis confirmed that older age, longer surgery duration, increased intraoperative blood loss, smoking, and diabetes are independent risk factors for postoperative infection (Table 3, Fig. 2).

Table 3.

Univariate and multivariate regression analyses of prognosis in the presence or absence of infection.

Norms	Univariate analysis				Multivariate analysis
	P	OR	95% CI		P	OR	95% CI
			Lower limit	Limit			Lower limit	Limit
Age	<.001	1.084	1.039	1.130	.006	1.073	1.020	1.129
Operation time	<.001	1.013	1.008	1.018	.009	1.008	1.002	1.013
Intraoperative blood loss	<.001	1.001	1.001	1.002	.030	1.001	1.000	1.002
Preoperative CRP	.041	1.033	1.001	1.065
Smoking	(Reference) no
Yes	.017	2.758	1.196	6.358	.013	3.149	1.278	7.758
Diabetes	(Reference) no
Yes	<.001	4.446	1.929	10.247	.005	3.914	1.513	10.127

CRP = C-reactive protein.

Figure 2.

Nomogram of infection.

3.4. Predictive efficiency analysis for postoperative infection risk

The ROC curve analysis was conducted using postoperative infection status as the state variable, with indicators showing differences in the multivariate regression serving as test variables. The results demonstrated that age’s area under the curve (AUC) for predicting postoperative infection was 0.754 (range: 0.658–0.849) (P < .001), with a specificity of 0.826, sensitivity of 0.565, and Youden Index of 0.391, identifying the optimal cutoff value as 67.5 years. The AUC for surgical duration was 0.743 (range: 0.627–0.860) (P < .001), with a specificity of 0.757, sensitivity of 0.696, Youden Index of 0.453, and the optimal cutoff value as 156 minutes. Intraoperative blood loss showed an AUC of 0.692 (range: 0.569–0.814) (P = .002), with a specificity of 0.806, sensitivity of 0.565, Youden Index of 0.371, and an optimal cutoff value of 475 mL. The AUC for diabetes as a predictive factor was 0.654 (range: 0.528–0.780) (P = .012), with a specificity of 0.829, sensitivity of 0.478, and Youden Index of 0.307. Smoking showed an AUC of 0.622 (range: 0.503–0.741) (P = .045), with a specificity of 0.680, sensitivity of 0.565, and Youden Index of 0.245 (Table 4, Fig. 3).

Table 4.

Risk prediction efficacy analysis for infection.

Norms	Cutoff value	AUC	95% CI		P	Specificity	Sensitivity	Youden Index
			Lower limit	Limit
Age	67.5	0.754	0.658	0.849	<.001	0.826	0.565	0.391
Operation time	156	0.743	0.627	0.860	<.001	0.757	0.696	0.453
Intraoperative blood loss	475	0.692	0.569	0.814	.002	0.806	0.565	0.371
Diabetes	–	0.654	0.528	0.780	.012	0.829	0.478	0.307
Smoking	–	0.622	0.503	0.741	.045	0.680	0.565	0.245

AUC = area under the curve.

Figure 3.

Risk prediction effectiveness ROC curve. ROC = receiver operating characteristic.

4. Discussion

Degenerative lumbar spine diseases, including lumbar disc herniation, spinal stenosis, and spondylolisthesis, lead to back pain and sensory-motor dysfunction in the lower limbs. TLIF is one of the primary methods for treating these conditions. There are minimally invasive TLIF and conventional O-TLIF. The former has the ability to minimize soft tissue and muscle damage compared to the extensive surgical exposure required for the latter, but O-TLIF is more suitable for dural adhesions, lumbar spondylolisthesis of more than II degree, spondylolisthesis, multisegmental surgery and fusion of the L5–S1. In this study, we analyzed the risk factors for incision site infections after performing O-TLIF. SSIs remain a common complication in orthopedic surgery, complicating healthcare delivery, prolonging hospital stays, increasing costs, and leading to more complications, even death. Our study analyzing factors related to SSI following O-TLIF found that age, smoking, surgery duration, diabetes, and intraoperative blood loss are its independent risk factors.

While some literature has explored the relationship between age and SSIs post-surgery,^[11,12] specific data on SSIs after lumbar surgery remain scarce. Age is often cited as a limiting factor for O-TLIF surgery due to studies suggesting that older patients have more comorbidities and because of concerns about increased risk of perioperative complications or poorer surgical outcomes. However, more studies have found that older patients (≥65) undergoing O-TLIF have not only similar perioperative outcomes but also the same improvements in pain, disability, and quality of life compared with younger patients (<65 years).^[13] Our multivariate analysis of 1000 patients who experienced SSI after lumbar surgery found that age (according to the ROC curve: over 68 years, increased infection likelihood) is a significant risk factor for post-lumbar surgery SSI. This finding helps provide a basis for future targeted preventive measures. Elderly patients often have comorbidities like cardiopulmonary diseases, fluid and electrolyte imbalances, liver and kidney function abnormalities. In the study by Peter Bischoff et al,^[14] the incidence of postoperative infection was significantly higher in patients over 70 compared to those under 50, possibly due to poorer general health and reduced host response to bacterial invasion associated with aging,^[15] and the loss of skin collagen with age increases the risk of infection at the surgical site.^[16] Age, an unmodifiable patient factor, can be managed with rational antibiotic use and enhanced postoperative care to prevent adverse events.

Our study found that smokers have a significantly higher risk of SSI post-lumbar surgery compared to nonsmokers, as smoking adversely affects tissue oxygenation and impairs the wound healing process and defense against pathogens.^[17] Research has shown that nicotine causes peripheral vasoconstriction and tissue hypoxia, damaging local angiogenesis and epithelialization.^[18] Smokers have poorer vascular conditions, negatively affecting wound healing. Martin et al categorized patients undergoing elective lumbar surgery into nonsmokers, preoperative quitters (12 months before surgery), and preoperative smokers. They found that preoperative smokers had a significantly higher risk of SSI compared to nonsmokers, while preoperative quitters had an increased risk but no significant difference from nonsmokers.^[19] Thus, supervising preoperative smoking cessation in elective surgery patients is essential.

In our study, surgery duration (according to the ROC curve: surgeries longer than 156 minutes significantly increase infection risk) was identified as a critical risk factor for SSI. Anis HK et al found a significant linear relationship between longer surgery duration and the occurrence of postoperative SSI, which remained significant even after adjusting for patient demographics, comorbidities, antibiotic use, and iatrogenic factors.^[20] This could be due to prolonged exposure of the incision and instruments during longer surgeries, increased bacterial counts in operating rooms due to staff movement,^[21] and extended tissue traction causing ischemic necrosis, thereby increasing the risk of wound contamination.^[22,23] Previous studies considered the number of surgical segments as a risk factor for SSI,^[24,25] but our results show that the number of surgical segments is not an independent risk factor for postoperative SSI, as it may reflect the surgery’s duration rather than directly affecting postoperative wound infection. Thus, intraoperative measures like relaxing wound tension and saline irrigation before closing the incision can effectively prevent bacterial colonization and may help reduce SSI risk.^[26,27] Apart from these preventive measures, staging the surgery into 2 phases is advisable when the duration is expected to exceed 5 hours^.[23]

Diabetes is a risk factor for postoperative wound infection in most published literature,^[5,28–30] and our study concurs, noting a significantly increased risk of SSI in diabetic patients. Research by Salazar et al^[31] found that prolonged hyperglycemia suppresses the function of inflammatory cells in the body, decreasing immunity and causing tissue damage. Diabetes-induced microvascular disease can lead to ischemia and hypoxia at the surgical site, delaying wound healing and increasing infection likelihood.^[32,33] Therefore, managing blood glucose perioperatively is crucial. Liu H et al conducted a study finding that preoperative subcutaneous insulin injections significantly reduce the incidence of post-lumbar surgery wound infection compared to other intervention groups, especially in patients with a long history of diabetes.^[34]

Our multivariate regression analysis identified intraoperative blood loss (according to the ROC curve: more than 475 minutes of blood loss significantly increases infection risk) as another risk factor. Tissue growth at the surgical site requires blood perfusion, and insufficient perfusion hinders tissue recovery, even leading to necrosis.^[35–37] Lasocki S et al found that anemia is closely related to SSI development.^[38] Nakanishi K et al reported that extensive allogeneic transfusion following blood loss suppresses a patient’s immune and coagulation functions, reducing wound healing capacity.^[39] Therefore, using hemostatic agents and meticulous surgical techniques to minimize blood loss are necessary.

In our study, the incidence of SSI post-lumbar surgery was 2.3%. Through this analysis of risk factors for postoperative wound infection, surgeons can implement reasonable interventions for high-risk patients preoperatively, such as stabilizing blood sugar, quitting smoking before surgery, reducing operation time, and prophylactic use of antibiotics, to decrease the incidence of postoperative SSI and promote rapid patient recovery, reducing unnecessary medical costs. While our study identified 5 independent risk factors for SSI, there are still limitations. Firstly, this is a retrospective study, which may be subject to selection and subjective biases. Secondly, surgeries were not performed by the same surgeon, possibly affecting the results due to varying levels of surgical experience. Lastly, our study only included data from 1000 patients, which, while sizable, is still not a vast sample size for risk factor prediction.

Author contributions

Formal analysis: Yanchun Xie.

Methodology: Kangen Han.

Resources: Anwu Xuan, Zhihao Zhang.

Software: Yin Hu, Shilei Tang.

Validation: Yuanhang Zhao.

Writing – original draft: Hailong Yu, Hongwen Gu.

Writing – review & editing: Hailong Yu, Hongwei Wang.