Does nuclear tissue infected with bacteria following disc herniations lead to Modic changes in the adjacent vertebrae?
To investigate the prevalence of infected herniated nucleus material in lumbar disc herniations and to determine if patients with an anaerobic infected disc are more likely to develop Modic change (MC) (bone oedema) in the adjacent vertebrae after the disc herniation. MCs (bone oedema) in vertebrae are observed in 6 % of the general population and in 35–40 % of people with low back pain. These changes are strongly associated with low back pain. There are probably a mechanical cause and an infective cause that causes MC. Several studies on nuclear tissue from herniated discs have demonstrated the presence of low virulent anaerobic microorganisms, predominantly Propionibacterium acnes, in 7–53 % of patients. At the time of a herniation these low virulent anaerobic bacteria may enter the disc and give rise to an insidious infection. Local inflammation in the adjacent bone may be a secondary effect due to cytokine and propionic acid production.
Patients undergoing primary surgery at a single spinal level for lumbar disc herniation with an MRI-confirmed lumbar disc herniation, where the annular fibres were penetrated by visible nuclear tissue, had the nucleus material removed. Stringent antiseptic sterile protocols were followed.
Sixty-one patients were included, mean age 46.4 years (SD 9.7), 27 % female. All patients were immunocompetent. No patient had received a previous epidural steroid injection or undergone previous back surgery. In total, microbiological cultures were positive in 28 (46 %) patients. Anaerobic cultures were positive in 26 (43 %) patients, and of these 4 (7 %) had dual microbial infections, containing both one aerobic and one anaerobic culture. No tissue specimens had more than two types of bacteria identified. Two (3 %) cultures only had aerobic bacteria isolated.
In the discs with a nucleus with anaerobic bacteria, 80 % developed new MC in the vertebrae adjacent to the previous disc herniation. In contrast, none of those with aerobic bacteria and only 44 % of patients with negative cultures developed new MC. The association between an anaerobic culture and new MCs is highly statistically significant (P = 0.0038), with an odds ratio of 5.60 (95 % CI 1.51–21.95).
These findings support the theory that the occurrence of MCs Type 1 in the vertebrae adjacent to a previously herniated disc may be due to oedema surrounding an infected disc. The discs infected with anaerobic bacteria were more likely (P < 0.0038) to develop MCs in the adjacent vertebrae than those in which no bacteria were found or those in which aerobic bacteria were found.
KeywordsBacterial infectionModic changesEndplate changesPropionibacterium acnesLumbar disc herniation
Modic changes (MCs) are bone oedema in vertebrae and have been shown to be both commonly observed and associated with low back pain [1, 2]. A recent systematic review showed that the prevalence for any type of MC in patients with chronic non-specific low back pain (CLBP) was 46 % as opposed to 6 % in the general population . A positive association between MC (bone oedema) and non-specific LBP was found in 70 % of studies with odds ratios ranging from 2.0 to 19.9 . The significance of these findings should not be underestimated as CLBP is universally acknowledged to be extremely difficult to reliably attribute to specific pathoanatomical causes .
MC are only visible on magnetic resonance imaging (MRI) and three types have been identified (Types 1, 2, and 3) . MC Type 1 consists of disruption and fissuring of the endplate, micro-fractures of the trabeculae, extracellular water and vascularised tissue within the adjacent marrow. MC Type 2 consists of disruption and fissuring of the endplate, micro-fractures of the trabeculae, and replacement of the haematopoietic elements of the marrow by yellow fat. MC Type 3 are sclerotic bone . The temporal development of MC is uncertain, but the time span is in years. Several studies have examined the reliability of reporting MC and all report excellent inter and intra-observer reliability [5, 6]. In comparative studies, MC has demonstrated a higher reliability than other MRI findings, such as disc herniations [7, 8].
There are two possible pathogenetic mechanisms resulting in MC. These are:
(1) A mechanical cause where degeneration of the disc results in the loss of soft nuclear material and reduced disc height and hydrostatic pressure. This increases the shear forces on the endplates with the result that micro-fractures may occur. The observed MC could be caused by oedema that is secondary to the fractures and subsequent inflammation, or the result of an inflammatory process associated with a toxic stimulus from the nucleus pulposus that permeates through the fractures .
(2) An infective cause in nuclear tissue removed under strict sterile conditions during surgery for lumbar herniated discs, 53 % of patients were found to have low virulent anaerobic microorganisms (Propionibacterium acnes and Corynebacterium propinquum) . This was in contrast to none of patients who had undergone surgery for other spinal disorders such as scoliosis . P. acnes are commonly found in hair follicles in the skin and in the oral cavity. They frequently invade the circulatory system during tooth brushing where they do not present an immediate health risk due to the aerobic environment of the blood stream [12–14]. When an intervertebral disc is herniated, nuclear material extrudes into the spinal canal. Within a short time, neocapillarisation begins in and around the extruded nucleus material [15–19] and inflammation occurs with an increased presence of macrophages [17–19]. As the avascular anaerobic disc provides an ideal environment for anaerobic bacteria, it is plausible that these low virulent anaerobic bacteria may enter the disc and give rise to a slowly developing infection. Local inflammation in the adjacent bone (MC Type 1) may be a secondary effect due to cytokine production or microbial metabolites (e.g. propionic acid) entering the vertebrae through normal disc nutrition. Studies have shown a dramatic (310 %) increase in the frequency of MC Type 1 following a lumbar disc herniation . The anaerobic bacteria are located in the anaerobic disc, and the MC may result from this adjacent infection. P. acnes is known from the skin to trigger an adjacent inflammatory response . P. acnes cannot multiply in the highly vasculated aerobic bone and are therefore not present where the MC occur as, shown by Wedderkopp et al. .
A small cohort study has also demonstrated a very positive response to treatment with antibiotics in patients with MC after a lumbar disc herniation . These findings lend credence to the theory that MC that develop after a lumbar disc herniation may be caused by a bacterial infection in some patients.
To investigate if herniated nucleus material from lumbar disc herniations is infected with bacteria.
To determine if patients with an anaerobic infected disc are more likely to develop MCs following a disc herniation as compared to patients with sterile discs or aerobic infections.
This study involved a cohort of patients undergoing primary surgery at a single spinal level for lumbar disc herniation. Patients were included if they were between 18 and 65 years of age and had an MRI-confirmed lumbar disc herniation, where the annular fibres were penetrated by visible nuclear tissue. Patients were excluded if they had received any antibiotic treatment within the previous 2 weeks. All patients had an MRI at baseline and 1–2 years after surgery.
Collection of biopsies
To avoid skin contamination, stringent antiseptic sterile protocols were followed, including, cleaning the skin in the operation field, preoperatively, for 2 min with 2 % chlorhexidine in 70 % isopropyl then allowing the solution to dry. The nucleus material was evacuated in five biopsies, each with a new set of sterile instruments. The five glass vials in which the nucleus material had been placed were immediately frozen at −80 °C. One high-dose 1.5 g cefuroxin (antibiotics) was administered intravenously after the biopsies were obtained to avoid inhibiting any bacteria present in the biopsies before the tissue was examined. The samples were kept at −80 °C until transported to Aston University in the special thermal transport boxes used for organ transport between hospitals. They were covered in frozen carbon dioxide (−80 °C).
The follow-up MRIs were all to be performed in the same scanner. An open low field 0.2 T, MRI unit with a body spine surface coil. The patients were placed in the supine position. Five sequences of localised images were taken: two coronal and three sagittal.
Patients were included regardless if they had Modic changes at baseline. Both baseline and follow-up MRIs were evaluated by the same consultant radiologist, who was an experienced research radiologist, blinded to the laboratory results. The type, size and volume were graded according to the Nordic Modic Protocol . At the follow-up MRI the occurrence of new MCs Type 1 or 2 at the level of the previous disc herniation was graded as a positive finding.
Detection, culturing and identification of bacteria
Microbiological examination of tissue samples
Five tissue samples were collected from each patient. Under sterile conditions, extracted tissue from each sample of the five was sectioned and finely ground using an individually packaged sterile, gamma-irradiated scalpel (Swann Morton) and a sterile, gamma-irradiated petri dish. Once the packaging was opened all scalpels were flamed before use as an extra precaution to ensure sterility. The five tissue samples from each patient, were then spread onto and embedded into Columbia blood agar plates (Oxoid, UK) which were then incubated under aerobic and anaerobic conditions for 7 days at 37 °C.
Microbial identification following anaerobic and aerobic culture growth
Following aerobic and anaerobic culture incubation, the resulting colonies were sub-cultured onto Columbia blood agar plates and incubated for 24 h at 37 °C in aerobic and anaerobic conditions. All colonies were investigated by Gram staining. Presumptive P. acnes colonies were identified following analytical profile index (API) biochemical analysis using the Rapid ID 32A kit (bioMerieux) and by polymerase chain reaction (PCR) amplification of 16S rDNA. Specific primers designed for amplification of P. acnes 16S rDNA were used to amplify a 600 bp region: forward primer 5′-GGGTTGTAAACCGCTTTCGCCT-3′ and reverse primer 5′-GGCACACCCATCTCTGAGCAC-3′ . Chromosomal DNA was extracted from single selected colonies by a rapid boil extraction method . PCR was performed in a 25-μl volume containing 19.8 μl of SDW, 2.5 μl of 10× PCR buffer (10 mM Tris HCl pH 8.3, 3.5 mM MgCl2, 25 mM KCl), 0.2 μl of each primer (25 pmol/μl), 0.2 μl of dNTPs (10 mM each nucleotide), 0.1 μl of Taq DNA polymerase (1.25 Units/μl) and 2 μl of template DNA. A positive control (with previously amplifiable P. acnes DNA) and a negative control (sterile water as template) were included. The following PCR conditions were used: initial 94 °C denature for 4 min; 35 cycles of 94 °C denature for 30 s, 54 °C annealing for 30 min, and 72 °C extension for 1 min; 72 °C extension for 4 min; 4 °C hold. A 2 % agarose gel containing 1 μg/ml of ethidium bromide was used to separate amplified fragments. Electrophoresis was performed in 1×TAE (40 mM Tris, 1 mM EDTA and 0.1 % (v/v) glacial acetic acid buffer at 100 V. Gram-positive cocci were further analysed using standard biochemical tests (oxidase and catalase) to identify presumptive staphylococci. In addition, latex agglutination for clumping factor/protein A was used to distinguish presumptive Staphylococcus aureus and coagulase-negative staphylococci.
Detection of bacterial DNA in tissue samples by PCR
To determine the presence of bacterial DNA in all tissue samples, including culture-negative samples, PCR amplification of bacterial 16S r DNA was undertaken. A sample of each tissue (approximately 50–100 mg wet weight) was aseptically transferred to a sterile 1.5 ml microcentrifuge tube. Each tissue sample was suspended in 200 μl of Tris–EDTA buffer, 100 μl proteinase K (10 mg/ml) and 240 μl sodium dodecyl sulphate (10 % w/v) and incubated for 48 h in a water bath at 45 °C. Following incubation, DNA was extracted from the lysed tissue samples with phenol/chloroform. Detection of microbial 16S rDNA was performed by PCR using broad range universal bacterial primers to amplify a 466-bp region: forward primer 5′-TCCTACGGGAGGCAGCAGT-3′ and reverse primer, 5′-GGACTACCAGGGTATCTAATCCTGTT-3′ . Amplification was performed as described above for the P. acnes identification. A positive control (using previously amplifiable P. acnes DNA) and negative control (sterile water as template) was run for every PCR performed. All samples giving a positive result with the broad range primers were also subjected to PCR using the P. acnes-specific primers as described above for identification of P. acnes colonies.
To determine the association between anaerobic culture and new Modic changes a Fisher’s exact test and odds ratio calculations were utilised. The statistical programme used was SPSS 13.
The study was approved by The Scientific Ethics Region of Southern Denmark no: VF-20060085. All patients signed a written informed consent.
Sixty-seven patients were included in the study. In three patients, it was not possible to evacuate nuclear material during surgery, and another three patients did not have an MRI taken at 1-year follow-up due to the long distance to The Spine Centre. Of the remaining 61 patients who were included in the study, the mean age was 46.4 years (SD 9.7) and 27 % were female. Of the 61 baseline MRIs, 50 were performed at the operating hospital. 11 came from 11 other hospital, and had a sufficient quality to evaluate the disc herniation and possible Modic changes. All follow-up, MRI were performed at The Back Research Center. The mean time between the two MRIs was 424 days (range 334–546).
All patients were immunocompetent. No patients had received a previous epidural steroid injection or undergone back surgery. None of the patients developed clinically evident post-operative discitis.
The distribution of bacteria in the positive cultures
Herniated discs (N = 61)
Of the herniated discs with positive microbiology (N = 28)
24 (40 %)
2 (3 %)
Gram-positive cocci (1 single)
4 (6 %)a
Gram-negative rod (1 single)
1 (1.5 %)
1 (1.5 %)b
52 % (46 %)d
Contingency table for the association between anaerobic culture and new Modic changes
New MCs at the site of the disc herniation
No new MCs
Positive anaerobic culture
Pure aerobic culture
The association between anaerobic culture and new Modic changes is shown as a contingency table in Table 2. A Fisher’s exact test shows this association to be highly statistically significant (P = 0.0038), with an odds ratio of 5.60 (95 % CI 1.51–21.95). If the contingency table is reconstructed using only those patients whose disc material was infected the association between anaerobic infection and new MCs is a relative risk of ‘infinity’ compared with those with aerobic infection, as no MCs developed in those with aerobic infection.
In this study, 46 % of the 61 patients with a lumbar disc herniation were found to have microorganisms present in extruded nuclear tissue. The microorganism most frequently cultured was the anaerobic bacterium, P. acnes. The discs from which anaerobic bacteria were isolated were stastistically significant (P < 0.0038) more likely to develop MC than those in which no bacteria or aerobic bacteria were detected. The findings of this study could be interpreted as a support to the theory that the occurence of Modic Changes (MCs) Type 1 in the vertebrae adjacent to the previously herniated disc might the oedema surrounding an infected disc.
This study confirms the original findings of Stirling et al. [10, 11] demonstrating that extruded nucleus material frequently has microorganisms present, a finding replicated in other studies. Corsia et al.  for example evacuated extruded disc material in 30 lumbar disc herniations and reported that 71 % had microorganisms present, 36 % with a staphylococcus and 18 % with P. acnes. In 30 cervical disc herniations they found that 59 % had microorganisms present, 37 % with P. acnes. Agarwal et al.  similarly cultured material excised from 52 patients treated with a single-level microdiscectomy for lumbar disc herniation. Ten (19 %) of the patients had microorganisms present and in 7 (70 %) of those, P. acnes was the sole microorganism isolated. Fritzell et al.  investigated 10 patients with a lumbar disc herniation without any clinical signs of an infection. Those biopsies were analysed with polymerase chain reaction (PCR) and in two patients (20 %) bacteria were identified, either Bacillus cereus or Citrobacter braakii/freundii. The authors expressed their surprise that a normally anaerobic tissue such as the disc had bacteria present in 20 % of their patients.
Ben-Gamlin et al.  expressed an alternative explanation of these findings. In their samples of evacuated nucleus material, they suspected that contamination was the reason for 2 (7 %) of the 30 patients having coagulase-negative staphylococci. The operations were performed under stringent sterile conditions and the samples were cultured both anaerobically and aerobically, but all patients received large doses of antibiotics (cefazolin) before the operation when the biopsy was later obtained. These high doses of antibiotics to which P. acnes is sensitive may partly explain the low isolation rate noted in this study.
Six studies have demonstrated the presence of bacteria in nucleus tissue evacuated from patients with lumbar disc herniation. In these studies the rates of isolation ranged from 7 to 53 %, which may reflect the extended length of incubation P. acnes requires for growth and the need for rapid processing following excision both of which can influence successful isolation. Different culture conditions like the growth medium and time will also affect the growth of P. acnes. In addition, further work by our group has demonstrated that patients with Modic type 1 changes in the adjecent disc following a lumbar disc herniation respond to long term antibiotic treatment with improvement in all measured outcome measures like pain, function, MRI, which could be interpreted as a possibility that the presence of these organisms is resulting in an infection .
The current study confirms the findings of other studies [10, 11, 27–29] where the microorganism most frequently found in the extruded nucleus material is P. acnes, accounting for approximately 80 % of the positive cases. The predominance of P. acnes might reflect the unusual environment in the disc where the lack of vascularity results in a very low oxygen tension and a low pH which provides ideal conditions for low virulent anaerobic bacteria to grow. These bacteria are unlikely to survive so well in the highly vasculated aerobic bone and are therefore not present where the MC occur which was confirmed by Wedderkopp .
The key element of concern is wether the detected bacteria found in the nuclear material from the herniated discs of these patients are indicative of an infection or possibly due to intra-operative skin contamination. Every possible precaution to avoid contamination was undertaken and the operation procedures were conducted under the strictest sterile conditions. If skin contamination was the cause of infection, a pattern of multiple skin bacteria cultures would be observed, and usually include coagulase-negative staphylococci. However, this was not the case with our findings where most were monocultures, a few were mixed cultures with two microorganisms, but there were no multi-cultures which would otherwise have been expected. In addition, Stirling et al.  showed in their study (using the same skin cleaning procedures as this study) that skin contamination was absent in control patients. In 207 lumbar disc herniation operations they found 37 % of the nucleus material had bacteria. This was in direct contrast to the 27 control patients (who underwent a similar spinal procedure but for scoliosis, trauma, or malignancy by the same surgical team using the same antiseptic procedures) from which no microorganisms were isolated. If the presence of microorganisms in the nuclear material was a result of skin contamination, then some of these 27 control patients would have been expected to have demonstrated positive cultures.
Evacuation of herniated disc material from the patients were not caused by a suspicion of a clinical relevant spinal infection. This study was done for research purpose only, the cultures and the data analysis was performed with this aim.
A further question which needs addressing is why do some patients develop MC when no microorganisms are present in their herniated nuclear tissue? In the paper by Albert et al. , at least two causal pathways for the development of new MCs are proposed. One theory is the infective pathway of the extruded nucleus material investigated in this study. But possibly the infection takes place later, after the nucleus material has been removed. After disc herniation surgery, MRI shows contrast enhancement along the surgical tract in all patients 6 weeks and 6 months after surgery . This enhancement is portraying hyper vascularisation, and it is possible that the bacteria might enter the disc at this later point. The other theory is the hypothesised biomechanical pathway. All patients in this study had a lumbar disc herniation where there was degeneration, loss of soft nuclear material, reduced disc height and hydrostatic pressure. These changes increase the shear forces on the endplates and micro-fractures may occur. The observed MC that developed in the absence of infection possibly due to the result of a biomechanical impact, reflecting oedema secondary to micro-fractures and subsequent inflammation, or a result of an inflammatory process from pro-inflammatory chemicals penetrating through the micro-fractures from the nucleus pulposus as proposed by Adams et al. [32, 33].
This study confirms the findings of five previous studies which suggested that up to half of patients undergoing surgery for a first time disc herniation have infected extruded nucleus tissue, which is normally sterile. The microorganism predominantly identified was the bacterium P. acnes). The findings of this study could be interpreted as supporting the theory that the occurrence of MCs Type 1 in the vertebrae adjacent to a previously herniated disc might be due to oedema surrounding an infected disc hence the discs that were infected with anaerobic bacteria were significantly more likely to develop MCs than those in which no bacteria were found.
Conflict of interest