Introduction. Degenerative disc disease results from mechanical alteration of the intervertebral disc. Biochemical modifications of the nucleus matrix are also incriminated. Furthermore, genetic predispositions as well as vascular factors have been advocated in the process of disc degeneration. A relationship between sciatica and Propionibacterium acnes has been described. However, it remains unclear if the hypothesis of a subclinical spondylodiscitis might play a role in the pathophysiology of degenerative disc disease. The purpose of this study was to analyze the possible presence of bacteria in lumbar discs of patients with degenerative disc disease. Methods. We prospectively analyzed the presence of bacteria in 83 patients (34 males and 49 females, average age 41 years) treated by lumbar disc replacement at L3-L4, L4-L5 or L5-S1. An intraoperative biopsy and microbiological culture were performed for each disc to determine if intradiscal bacteria were present. Great care was taken to avoid any source of contamination during the conditioning process of the biopsy. Microbiological results were compared to the magnetic resonance stages of disc degeneration according to the Pfirrmann and Modic classifications. Possible sources of previous iatrogenic disc contamination after discography or nucleotomy were analyzed. Results. The magnetic resonance stages were Pfirrmann IV or V, with Modic I signs in 32 and Modic II in 25 cases. A preoperative discography was performed in 49 patients, and 24 had previous nucleotomy. Germs were found in 40 discs, 43 cultures were steriles. The following bacteria were evidenced: Propionibacterium acnes 18, Staphylococcus coagulase negative 16, Staphylococcus aureus 3, Gram negative bacilli 3, Micrococcus 3, Corynebacterium 3, others 5. Ten biopsies presented several different germs. Bacteria were predominantly found in males (p=0.012). The mostly positive level was L4-L5 (p=0.075). Histological examination of 31 discs found inflammatory cells in 33 % of the biopsies with positive bacterial culture, versus 5 % of the sterile biopsies (p=0.038). There was no significant relationship between bacterial evidence and Modic sign. A preoperative discography or previous nucleotomies did not represent significant contamination sources. None of the patients presented clinical signs of infection. Conclusions. The finding of bacteria in 48 % of disc biopsies, presence of inflammatory cells at histological examination, the absence of responsibility of the discography as a factor of contamination, and the absence of clinical post-operative infection, defend the hypothesis of a low-grade spondylodiscitis which might play a role in the pathophysiology of degenerative disc disease. On the other hand, the presence of skin commensal bacteria, of ¼ of polymicrobial biopsies, and the fact that previous nucleotomy doesn't seem to be involved in inoculation, cannot allow to eliminate a contamination of the samples. Further studies are necessary to elucidate the responsability of intradiscal bacteria in degenerative disc disease. This could influence our treatment strategy of back pain, which could be based in the future on antibiotics