Postoperative spondilodiscitis

Abstract

Introduction

Postoperative spondylodiscitis is a primary infection of the nucleus pulposus with secondary involvement of the cartilaginous endplate and vertebral bone. Although uncommon, postoperative spondylodiscitis causes major morbidity and may be associated with serious long-term sequelae. Several risk factors had been identified, including immunosuppression, obesity, alcohol, smoking, diabetes and malnutrition.

Materials and methods

A review of the literature was done to analyse the diagnosis, treatment and prevention of postoperative spondylodiscitis.

Results

We found that the principles of conservative treatment are to establish an accurate microbiological diagnosis, treat with appropriate antibiotics, immobilise the spine, and closely monitor for spinal instability and neurological deterioration. The purpose of surgical treatment is to obtain multiple cultures of bone and soft tissue, perform a thorough debridement of infected tissue, decompress neural structures, and reconstruct the unstable spinal column with bone graft with or without concomitant instrumentation.

Conclusions

Appropriate management requires aggressive medical treatment and, at times, surgical intervention. If recognised early and treated appropriately, a full recovery can often be expected. Therefore, clinicians should be aware of the clinical presentation of such infections to improve patient outcome. A review of the literature was done to advance our understanding of the diagnosis, treatment, prevention and outcome of these infections.

Introduction

Postoperative spondylodiscitis is a rare primary infection of the nucleus pulposus with secondary involvement of the cartilaginous endplate and vertebral bone after spinal surgery [1, 2]. Two causes are possible, haematogenous spread, less frequently, or direct inoculation of virulent organisms during the surgery [1, 3, 4]. The aim of this article was to review the clinical features, operative management, outcome in patients and most important the way to diagnose and prevent it.

Incidence

The incidence of postoperative spondylodiscitis varies from 0.21–3.6% [6–9]. In most of the reports, the average incidence is 1% after simple discectomy [10]. The most common etiology of postoperative spondylodiscitis is bacterial infection [6, 11], but responsible pathogens include bacteria and fungi [12].

Predisposing factors

Several patient-related factors play a role in the pathogenesis of infection; some are modifiable, whereas others are not. Advanced age, developmental delay, immunosuppression, spinal trauma, and diabetes mellitus are some risk factors that are not modifiable. Obesity, smoking, indwelling catheters, malnutrition, and extended hospitalisation are modifiable risk factors. High rates of infection have been observed in elderly patients and spine trauma patients [13–14]. Obesity has also been quoted as a risk factor by many [15–17].

Malnutrition, objectively defined by a serum albumin level less than 3.5 g/dL, a total lymphocyte count less than 1,500–2,000 cells/mm³ and a serum transferrin level of 150 μg/dL is another documented risk factor.

Malignancy, chemotherapy, immune suppression, and diabetes mellitus have also been implicated as risk factors for postprocedural infections [8, 17, 18].

Operating through previously irradiated tissue or previously operated tissue has also shown an increased risk for surgical site infection [16, 19].

Etiology

Postoperative pyogenic spondylodiscitis

The leading cause in most of the cases is Staphylococcus aureus (60%) [10, 12] and Gram-negative organisms [12]. In 2011, Li et al. [20] reported a retrospective study on 34 patients treated for postoperative intervertebral discitis including ten cases after lumbar discectomy, 21 cases after percutaneous lumbar discectomy, two cases after percutaneous cervical discectomy and one case of C5-C6 after percutaneous cervical nucleoplasty. Most of the authors consider that the main cause of these infections is an inoculation during surgery and less frequently a haematogenous contamination.

Some cases have been described in postoperative discitis due to Propionibacterium acnes, a common skin anaerobe. In the case reported by Harris et al. [21], infection occurred seven weeks the after the index procedure. The patient underwent surgery with irrigation and debridement combined with antibiotic therapy. Two months later, there was an increased deformity by a destructive osteomyelitis with neurological dysfunction that required surgery. The patient underwent anterior lumbar interbody fusion at L5-S1 and posterior spinal fusion at L4-S2. The patient ended with neurological damage.

A recent retrospective study of Hsieh et al. [22] with 6,120 cases of instrumented posterolateral fusion, reported 11 patients with a postoperative anterior spondylodiscitis. The organisms found were: methicillin-resistant coagulase-negative staphylococci in two cases, methicillin-sensitive coagulase-negative staphylococci in two cases, methicillin-resistant Staphylococcus aureus in two cases, methicillin-sensitive Staphylococcus aureus in one case and Serratia marcescens in two cases.

Carmouche et al. [23] reported that after instrumented posterior lumbar interbody fusion, the patient developed an epidural abscess and discitis, and the surgical cultures were positive for Staphylococcus aureus.

Kulkarni et al. [24] reported a case of adjacent level discitis after anterior cervical discectomy and fusion and the perioperative biopsy was positive for Serratia marcescens.

Postoperative tuberculosis spondilodiscitis

In the literature, we found three cases of tuberculosis disc infections after surgery. Two of these reported cases were diagnosed three months after the index procedure [25, 26]. Lotfinia et al. [15] reported a case in 2010 of late-onset postdiscectomy tuberculosis at the same lumbar level eight years after simple L4-L5 discectomy. The features of this disc tuberculosis infection are the difficulty of diagnosis, with nonspecific clinical signs and culture and the duration of the treatment. The TB infection comes from haematogenous infection, not detectable in most cases [27], that may come from an indolent pulmonary focus [27]. In the case reported by Jeon et al. [25], the patient underwent two operations and several months of antibiotic treatment before the diagnosis. In the last case reported by Kaplan et al. [26], the patient underwent three operations and inadequate antibiotic treatment before the diagnosis.

Postoperative fungal spondylodiscitis

Postoperative fungal spondylodiscitis is a rare cause of infection. In the literature, 14 cases of postoperative fungal spondylodiscitis have been described [28, 29], nine caused by molds and five by Candida species. The median age was 49.4 (range 19–64) years, consisting of eight men and six women. There were three cases (21%) after laminectomy and discectomy, two cases (14%) of laminectomy, four cases (31%) of discectomy and one case (1%) of nucleolysis. The most common organism was Aspergillus, in six cases (43%), then Candida in five cases, one case of Scedosporium prolificans, one case of Rhizopus rhizopodoformis and one case of Trichosporon asahii. According to the level of the infection, there was one at L3-L4, nine at L4-L5, and two at L5-S1. The most common symptom was back pain, found in 13 cases (93%), fever found in three cases, neurological deficit in two cases and one paraplegia. The clinical symptoms appeared at an average six weeks after the primary procedure (1-11). MRI showed osteomyelitis and/or discitis in all cases, complicated by paravertebral abscess in three cases and by epidural abscess in four cases. All patients received antifungal drugs. Thirteen patients underwent a surgical debridement, six had only a debridement, four underwent a laminectomy, one patient underwent a laminectomy and posterior fusion with instrumentation, and two patients underwent an anterior fusion with autograft and posterior fusion with instrumentation. One patient had no surgery. Ten patients had a complete recovery, one patient retained numbness of the right lower extremity. Two relapsed after the treatment, one patient four weeks later, after having been treated with intravenous and oral antifungal medication, and bed rest. The other, three months later, underwent a further debridement and antifungal medication. One of the patients underwent revision surgery 11 months later because of persistent lower back pain; he had hardware removal with improvement of symptoms after the procedure.

The most probable source of infection was direct invasion of fungus into the disc space during the primary procedure. Three of the five reported cases of Candida infection were patients with broken artificial nails removal with injury to the skin [30].

Preventive nosocomial measures are important to avoid these infections. Garcia et al. [30] highlight the importance of ventilation filters in the operating theatre. The conventional ventilation and filters in operating rooms remove airborne particles over 5 μm in size and most mold spores are larger than this. Fungal infection diagnosis must be considered when cultures from the biopsy site are reported negative for bacteria and the symptoms persist despite antibiotic treatment. It is recommended in postoperative spondylodiscitis to look for these fungi and mycobacterium in addition to bacteriological routine and even make a surgical biopsy if the diagnosis is strongly suspected of mold infection and the needle biopsies are negative [30]. Radiological examinations are not specific to these infections. These fungal infections usually had a very good prognosis after surgical debridement and antifungal therapy.

Clinical findings

The typical postoperative course of discitis consists of initial resolution of symptoms, with subsequent return of low back pain approximately six weeks postoperatively [31, 32]. In the series of 17 patients of Bavinzski et al. [33], all patients had localised back pain with muscle spasms and were aggravated by movement. Six patients had fever. Most of the patients became symptomatic within four weeks of their discectomies. The length of the clinical course before definitive diagnosis of postoperative discitis and reoperation ranged from three days to 15 months (mean 66.5). The positive straight leg-raising signs were seen in 13 patients. The pain might radiate to the hip, scrotum, perineum abdomen, and lower limbs. One patient had emergency surgery for a wound infection and purulent discharge.

In the reported case of postoperative infection with Propionibacterium acnes by Harris et al. [21], clinical symptoms are quieter; the patient was afebrile with a moderately elevated temperature and tenderness on palpation of the problem area. This appears between three weeks and three months postoperatively. But despite appropriate antibiotic therapy and initial surgical treatment, osteomyelitis progressed and the patient had a neurological deficit with lower extremity weakness requiring the use of a walker.

Laboratory findings

Elevated ESR and CRP are extremely sensitive [32, 34]. In the series of 17 patients of Bavinzski et al. [33], CRP values were increased before reoperation in 16 patients, and white blood cells were increased in only six patients associated with a shift to the left in the differential blood count. Only three blood cultures were positive. The main micro-organisms were: Staphylococcus epidermidis (5), Staphylococcus aureus (4), Staphylococcus coagulase negative- (1) and gram-positive cocci (5). In two patients, no bacteria could be identified. In 13 patients, histological examination found an inflammatory process.

In 2011, Li et al. [20] reported a retrospective study on 31 patients treated for postoperative intervertebral discitis by percutaneous discectomy and drainage. In 17 cases the cultures of disc and bone tissue biopsy were negative. The remaining 14 had positive cultures, with Staphylococcus aureus in seven cases, Staphylococcus epidermidis in two cases, Escherichia coli in two cases, Klebsiella in two cases and one case of Pseudomonas cepacia. These 17 negative results may be explained by inadequate amount of specimen, sampling error or previous antibiotic therapy. If the blood cultures are negative, the recommendation is to do an open biopsy [31]. In the series of Li [20], the biopsy of microbiological yield was 45.2%. And for the others patients with negative cultures (54.8%), they believe the no-infection discitis may have been caused by a chemical or aseptic process, or the pre-biopsy administration of antibiotics [35].

The traditional technique [36] to make the diagnosis of these infections is needle biopsy. Using a Mazabraud trocar, Fouquet et al. [35] obtained bacteriological diagnosis in nine (36%) of 25 patients. Other methods have been described, such as CT guided biopsies. Enoch et al. [37] reported a retrospective study of CT-scan guided biopsy in 103 patients. They found 27 cases of discitis and paravertebral abscess and the cultures were positive in nine of 25 (36%) samples. Staatz et al. [38] reported a series of biopsy CT-guided, percutaneous and catheter drainage, with 16 positive cultures (76%) for 21 patients.

In 2011, Li et al [20] described the percutaneous discectomy technique, which allows biopsy, decompression and irrigation of the disc space. They suggest this management as an alternative to open surgery [39–41].

Polymerase chain reaction is also another method of diagnosis with high sensitivity, specificity and, positive and negative predictive values [42–45]. Fenollar et al. evaluated the usefulness of broad-range PCR to improve the aetiological diagnosis of osteoarticular infections. For them, broad-range PCR followed by sequencing is the modality of choice in cases where the cultures are negative with positive clinical findings. This technique is useful for the diagnosis of patients taking antibiotics. Several authors have also used it for the diagnosis of Mycobacterium tuberculosis [42–45].

Imaging studies

MRI remains the most sensitive and specific imaging modality [46, 2]. In the series of 17 patients of Bavinzski et al. [33], the radiological examination performed for diagnosis was the MRI. In other series the MRI was the imaging modality of choice in the diagnosis of postoperative discitis with a reported sensitivity and specificity of 93% and 97%, respectively [46, 2]. MRI findings of postoperative discitis include a decreased signal intensity on T1-weighted images and an increased signal intensity on T2-weighted images in the disc space, secondary to oedema from the inflammation and infection (Modic I changes). After injection of gadolinium, there is a decreased signal intensity in the adjacent bone marrow on T1-weighted images.

In 2011, Li et al. [20] reported a retrospective study on 34 patients with postoperative intervertebral discitis. They evaluated their radiographic results and found 29 cases of disc space narrowing, 14 cases with destructive and sclerotic vertebral body changes with narrowing of the disc space.They evaluated their radiographic results and found 29 cases of disc space narrowing, 14 cases with destructive and sclerotic vertebral body changes with narrowing of the disc space. Nielsen et al. [5] reported a large series of 53 patients with postoperative discitis; in 17% of cases the interspace was fused. In Wirtz et al. [47] they found 65.5% spontaneous fusion.

In the study of Hsieh et al. [22], all patients had loosening of the pedicle screws. Kulkarni et al. [24] reported a case of adjacent level discitis after anterior cervical discectomy and fusion, and the radiographs showed segmental C4-C5 kyphosis, decreased disc height at this level with erosion of end-plates, halo around the screws and enlarged pre-vertebral soft tissue shadow. MRI showed discitis at the C4-C5 level with prevertebral and epidural abscesses from C4-C6.

Reviewing other imaging studies, Ohtori et al. analysed the utility of the 18F-fluorodeoxyglucose-PET for patients with suspected spondylitis showing Modic change in a prospective cohort study of 312 patients. Eighteen patients were suspected of having pyogenic spondylitis. In conclusion, the rate of detecting spondylodiscitis infection was very high if FDG-PET was additionally used. FDG-PET is recommended to distinguish between common Modic change and spinal infection.

Prevention

It is very important to note that the only factor found in the literature to prevent infections was pre- and intra-operative antibiotic prophylaxis [32, 34]. Brown et al. recommend a first- or second-generation cephalosporin, or a glycopeptide plus gentamicin for patients who are allergic to cephalosporins or who are colonised with methicillin-resistant Staphylococcus aureus, and they also recommend it in cases of prolonged operative procedures. In operations with long incision and exposure they recommend the use of frequent (e.g., every 15 minutes) saline irrigation with or without providone iodine to reduce the wound contamination.

Based on their review of literature, Brown et al. found that closed suction drains do not increase the risk of postoperative wound infection or the incidence of haematoma formation in patients undergoing simple lumbar laminectomy.

Treatment

Postoperative discitis after discectomy

In 2003, Bavinzski et al. [33] reported a series of 17 patients with septic postoperative spondylodiscitis, who were treated by early microsurgical removal of the infected necrotic tissue, irrigation of the disc space, and application of a closed suction-irrigation system, specific antibiotic therapy and early mobilisation in a light cast corset. Antibiotics were first administered IV for three to four weeks and two to three months orally. With this management, they obtained excellent or good clinical long-term results in 82% of the patients whereas 18% had poor results. All patients had a pain relief immediately after the surgery except one, who had a third procedure before improvement.

In 2011, Li et al. [20] reported a retrospective study of 34 patients to evaluate the effect of percutaneous discectomy and drainage for postoperative intervertebral discitis. Elevated CRP and ESR values returned to normal range within three to eight weeks.

In two of three reported cases of tuberculosis postdiscectomy patients underwent antibiotic therapy and surgical treatment. In most cases of tuberculosis spondylodiscitis, the treatment is medical with appropriate antibiotic therapy even in cases of abscess [48]. If the patient has neurological deficit, an abscess that did not resolve with antibiotics, and spinal instability [49, 50], open surgery is recommended [51, 48].

Postoperative discitis after instrumentation and interbody fusion

In the Carmouche series [23] of postoperative discitis after instrumentation and interbody fusion with an epidural abscess, the patient underwent two procedures to eradicate the infection, the first was surgical debridement and epidural abscess drainage and because of the worsening clinical symptoms, underwent repeat debridement and the removal of his interbody cage and pedicle screw instrumentation four months after index surgery. At ten-month follow-up, the posterolateral arthrodesis was healed and his back pain had improved compared to preoperative levels.

Kulkarni et al. [24] reported a case of adjacent level discitis after anterior cervical discectomy and fusion. The patient underwent surgical treatment with debridement, irrigation, and removal of hardware, and reconstruction of infected level by tricortical iliac crest autograft; the cages were left intact. He had IV antibiotic treatment for three weeks and oral medication for three weeks. The neurological recovery was complete and at eight-month-follow-up the fusions were sound.

Clinical outcome

In the series of 17 patients of Bavinzski et al. [33], the mean follow-up was 4.7 years (range, 8 months to 8.6 years). Four patients had an excellent clinical result, ten patients had good results and three had poor results. All patients who had a postoperative MRI showed resolution of the infectious process.

In 2011, Li et al. [20] reported a retrospective study on 34 patients to evaluate the effect of percutaneous discectomy and drainage for postoperative intervertebral discitis. The average follow-up was 6.4 years (range, 2–10). The VAS scores demonstrated statistically significant improvement after percutaneous discectomy and drainage.

Iversen et al. [52] reported that the patient that had a postoperative discitis had a significantly higher incidence of chronic low back pain.

In the reported case of postoperative infection with Propionibacterium acnes by Harris et al. [21], the patient had a neurological deficit. The patient needed bilateral articulating ankle foot orthoses and a walker. In the retrospective study of Hsieh et al. [22], all patients with a postoperative anterior spondylodiscitis after pedicle screw instrumentation improved their clinical outcomes, VAS and Oswestry Disability Index.

Conclusion

Postoperative spondylodiscitis is rare but can happen and the most important factor to prevent it according to the literature is pre- and intra-operative antibiotic prophylaxis [32, 34]. Thus, the diagnosis and the treatment are very important. We also have to consider the risk factors involved, which include immunosuppression, obesity, alcohol, smoking, diabetes and malnutrition [10]. In conclusion, early diagnosis can decrease the morbidity and mortality of this disease [53, 54].