Comparison Between Pyogenic Postoperative and Native Vertebral Osteomyelitis : Clinical Features, Curative Effect and Analysis of Prognostic Factors


 BackgroundPyogenic spondylodiscitis(PS) is a potentially life-threatening infection burdened with high morbidity rates. Despite the rising incidence, the proper diagnosis and treatment of PS are still controversial. Postoperative Vertebral Osteomyelitis(PVO) is a clinical challenge, for there were few reports about the treatment results of PVO before, and further more few studies have compared PVO with native vertebral osteomyelitis(NVO). The purpose of this study was to compare and describe the microbiology, clinical characteristics, treatment and curative effect between PVO and NVO, and analyze the prognostic factors as well. MethodsThe clinical data of 52 patients with pyogenic spondylitis admitted to the Spine Surgery Department of the First Affiliated Hospital of Xinjiang Medical University from January 2010 to December 2019 were retrospectively analyzed. There were 30 patients in native vertebral osteomyelitis (NVO) group, including 18 males and 12 females, with an average age of 50.47 ± 20.45 years old (aged from 15 to 73); 22 patients in postoperative vertebral osteomyelitis (PVO) group, including 13 males and 9 females, with an average age of 51.45 ± 16.97 years old (aged from 14 to 73). In Group NVO, 23 cases (76.7%) were located in lumbar vertebrae, 5 cases (16.7%) in thoracic vertebrae and 2 cases (6.7%) in cervical vertebrae; in Group PVO, 16 cases (72.7%) in lumbar vertebrae and 6 cases (27.3%) in thoracic vertebrae. 29 patients had had neurological dysfunction before surgery was taken. There were 26 cases of grade D (16 cases in Group NVO and 10 cases in Group PVO) and 3 cases of grade C (1 case in Group NVO and 2 cases in Group PVO), following the instructions of American Spinal Injury Association (Asia) neurological function classification. All patients were given bed rest, nutritional support and antibiotic therapy; surgical treatment for patients with poor outcomes or aggravated symptoms. Patients were followed up at 1, 3, 6 and 12 months after surgery, including leukocyte count, ESR and CRP, X-ray, CT three-dimensional reconstruction and MRI were performed. The changes of visual analogue scale (VAS) and Asia neurological function classification were observed to evaluate the clinical efficacy simultaneously. ResultsAll patients were followed up for 12-24 months. Till the last follow-up, 3 patients in Group NVO recurred, the recurrence rate was 10% (3 / 30), 9 patients in Group PVO recurred, the recurrence rate was 40.1% (9 / 22), the recurrence rate of Group PVO was higher than that of Group NVO, the difference was statistically significant (P = 0.009). Both groups were treated with intravenous and oral antibiotics, and the time of antibiotic treatment in Group PVO was longer than that in Group NVO, however the difference was not statistically significant (P = 0.094, P = 0.062). Among 44 patients with spinal internal fixation, 13.6% (1 NVO, 5 PVO) had recurrent infection after internal fixation. Therefore, we took re-operation to remove the internal fixator for infection control, patients recovered after conservative treatment such as immobilization and systemic anti infection. The numerical value of leukocytes, C-reactive protein, ESR and VAS scores of the two groups were significantly lower than those before surgery, the difference was statistically significant (P < 0.001). In Group NVO, 16 cases recovered from Asia grade D to grade E, 1 case from grade C to grade D; 10 cases in Group PVO recovered from grade D to grade E and 2 cases recovered from grade C to grade D. There was no significant difference between these two groups (P > 0.05). By univariate analysis, multiple vertebral involvement and abscess formation (P = 0.003, P = 0.025) were significantly associated with PS recurrence; there was a tendency for PS recurrence among microbial infection (OR = 1.889), spinal prosthesis (OR = 7.083) and allogenic bone (OR = 2.032), yet not obvious. By multivariate analysis, we found that multiple vertebral involvement (OR= 12.656, 95% CI: 1.536-104.303, P = 0.018) was a risk factor for PS recurrence. ConclusionThe treatment of PVO is more challenging than NVO, especially in the cases of spinal implant infection. Although the antibiotic treatment time of PVO is longer than that of NVO, the recurrence rate of PVO is higher. Longer antibiotic therapy and, if necessary, surgical debridement or removal of implants are important approaches to successful treatment of PVO.

The two common types of PS are postoperative vertebral osteomyelitis (PVO) and native vertebral osteomyelitis (NVO). Although both types are caused by purulent bacteria, but the pathogenesis, clinical manifestations, treatment methods, treatment time, prognosis, complications and recurrence rate of the two spinal infections are different. Many scholars have begun to pay attention to the treatment and efficacy of PS in recent years, and related literature has also been reported, however there are few reports comparing PVO and NVO.
The purpose of this study is to compare and describe the microbiology, clinical characteristics, treatment and efficacy of PVO and NVO, and analyze the related factors from pathogenesis to prognosis, so as to provide scientific basis for making standard diagnosis and treatment plan of PVO and NVO respectively.

General information
The clinical data of patients with PS admitted to Spine Surgery Department, the First

Preoperative management
Bacterial culture and drug sensitivity tests were performed before antibiotics were administered to all patients after admission. Blood culture and drug sensitivity tests were performed for patients with body temperature over 38℃. Patients with positive bacterial culture were treated with sensitive antibiotics according to drug sensitivity test; patients with negative bacterial culture were treated with empirical broad-spectrum antibiotics and monitored more closely. The total course of intravenous antibiotics was more than 6 weeks.
Surgical treatment is required when neurological dysfunction, conservative treatment failure, spinal instability (spinal instability includes segmental kyphosis > 15°, vertebral collapse > 50%, and displacement > 5mm.), huge paravertebral abscess or epidural abscess, and severe pain occur [8]. In Group NVO, 2 cases were treated with antibiotics only, the other  Table 2).The removed lesion tissues were sent for pathological examination, bacterial culture and drug sensitivity tests.

Follow-up evaluation and statistical analysis
All patients were followed up at 1,3,6 and 12 months after operation, including clinical symptoms, leucocyte count, ESR and CRP to evaluate infection control, VAS score to evaluate pain improvement, Asia neurological function classification to evaluate neurological function improvement,etc.; at the same time, X-ray and three-dimensional CT scanning were performed to evaluate the lesion repair, bone graft fusion and internal fixation, and MRI scanning was performed if necessary to detect the healing of the diseased vertebrae and the condition of the spinal cord.
All statistical analyses were performed with SPSS 26.0 statistical software. The measurement data accorded with normal distribution or approximate normal distribution(x ± s ), the differences between the two groups were compared by t-test, the partial distribution data was expressed by M(P25, p75), the differences between the two groups were compared by Mann Whitney U test, the value of count data cases was expressed as(%), the differences between the two groups were compared by C2-test or Fisher exact test. Univariate and multivariate logistic regression analyses were used to identify risk factors associated with treatment failure or relapse. Variables with P < 0.10 after univariate analysis were included in multivariate analysis. Significance tests were two-tailed, and P-values< 0.05 indicated statistical significance.

Results
60% patients in Group NVO and 59.1% in Group PVO were males. The proportion of male patients in both groups was higher than that of female patients, but there was no significant difference between the two groups in gender (P > 0.05). Diabetes was the most common comorbidity (Group NVO 23.3%, Group PVO 36.4%), more common in Group PVO, but the difference was not statistically significant (P = 0.306) ( Histopathological examination showed that fibrous tissue was proliferative and degenerative, with infiltration of plasma cells, lymphocytes, neutrophils and other acute and chronic inflammatory cells, and thin yellow pus with necrosis could be found.
Multiple vertebral involvement was more common in Group PVO, and the difference was statistically significant (P=0.046). The proportion of patients with epidural abscess and paraspinal abscess formation was not significantly different between the two groups (P > 0.05).  1) The scores of white blood cell, C-reactive protein, ESR and VAS in the two groups were significantly lower than those before operation at 1, 3, 6 months and the last follow-up, and the differences were statistically significant (P < 0.001) (  was a risk factor for PS recurrence.

Discussion
PS refers to the infection of vertebral body, intervertebral disc and soft tissue around vertebral body caused by pyogenic bacteria [9], it takes up 4% of all kinds of osteomyelitis, including vertebral osteomyelitis, discitis and epidural abscess [10]. Improper or delayed treatment of PS may lead to serious complications. Its incidence is on the rise, and although there has been great progress in the understanding of this disease, doctors are still facing great challenges about the diagnosis and treatment [11]. Despite surgery and antibiotic treatment, the clinical outcomes yet remain poor [12].
In our study, as a part of non-specific clinical manifestations, pain was the most common symptom. All patients showed local pain (52 / 52, accounting for 100%), which was similar to the results of Hopkinson N et al. [13].In Group NVO, the local symptoms of low back pain were more severe than the patients in Group PVO, the VAS score of Group NVO was (7.10 ± 1.30), which was higher than that of Group PVO (5.73 ± 0.985) before surgery.
Most studies have shown that white blood cells and neutrophils are not the sensitive factors for diagnosing PS. Yoon et al. [14] pointed out that only 42.6% -81.3% patients with suppurative spondylitis had elevated leucocyte count. Domestic studies such as Polat [15] and others showed that 58% patients with PS had elevated neutrophil and leucocyte count. On the contrary, erythrocyte sedimentation rate(ESR) and C-reactive protein(CRP) were more sensitive, the sensitivity was 98% and 100% respectively, but the specificity of ESR and CRP was poor, which can be used to observe the progress of disease and evaluate the curative effect [16]. Like the results reported by Yoon and Polat et al., 26.7% patients in Group NVO and 40.1% patients in Group PVO increased the leucocyte count in this study. Although the leucocyte count in Group PVO increased more significantly, the difference was not statistically significant(P > 0.05), and all patients showed increasing ESR and CRP, so we think that ESR and CRP detection is more sensitive than leucocyte count. Although the specificity of CRP is low, Jean et al. [17] found that CRP value over 63 mg/L was signifcantly associated with shorter diagnostic delay. Therefore, the association of spinal pain and high value of CRP could increase the suspicion of PS and lead to early diagnosis.
Identification of pathogens is the cornerstone of PS diagnosis and treatment. CT-guided biopsy is often needed for patients without surgical indications. The specificity of CT-guided biopsy was 99%, and yet the sensitivity was 52% -91% [18]. Spira et al. [19] reported that the positive rate of CT-guided biopsy in 34 patients with PS was 82.3%. Marschall et al. [20] showed that the positive rate of open biopsy culture in PS patients was 91%, while that of percutaneous biopsy was only 53%.The common gram-positive pathogens include was longer than that in Group NVO, the difference was not statistically significant(P = 0.094).
In our patients with recurrence related risk factors (such as abscess in the surrounding tissue of the spine, multiple vertebral body involvement, etc.), the antibiotic treatment time was longer (＞8 weeks), which was similar to the results of Park et al. [26].
Systemic support therapy, antibiotic therapy and local immobilization are the preferred choices for PS. But surgical treatment is still recommended for patients with neurological impairment, conservative treatment failure, spinal instability, huge paravertebral abscess or epidural abscess, and severe pain. Spinal instability includes segmental kyphosis > 15°, vertebral collapse > 50%, and displacement > 5mm [8].
At present, the choice of surgical approach for PS is still controversial. The anterior approach can make the infected part fully exposed, can directly and clearly see the lesion tissue, and then can fully and thoroughly remove the necrotic and infected tissue. At the same time, it can fully drain the concurrent abscess, so as to prevent the residual infection focus and bacterial planting. However, because the anterior internal fixation is not strong enough and the risk of infection is high, it is easy to lead to surgical failure. There are also some scholars through the posterior approach debridement and get better results, nerve root decompression, deformity correction is suitable for posterior surgery. According to this point of view, the combination of anterior and posterior surgery is the best choice for PS surgery.
However, due to the large trauma and long operation time, especially for the elderly patients, it is not often used, which is still controversial. Since the anterior approach is associated with higher incidence rate and mortality, recent literature has shown that the posterior approach can adequately solve the infection problem [27] .In this study, 17

Availability of data and materials
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.