Lumbar Spondylodiscitis Caused by Propionibacterium acnes After Epidural Obstetric Analgesia

Propionibacterium acnes is a gram-positive anaerobic organism that is a usual inhabitant of normal skin, and like many other organisms, it is capable of producing spondylodiscitis. Disk infection by P. acnes has been reported in immunocompetent individuals by direct contamination at the site of puncture, after the implant of foreign materials, or by hematogenous spread ^(1,2). We report a case of iatrogenic spondylodiscitis by P. acnes that occurred after the placement of an epidural catheter for analgesia in a healthy parturient. Infection of intervertebral discs and of the adjacent vertebrae by P. acnes constitutes a serious condition, which has not been previously reported complicating epidural obstetric analgesia.

Case Report

A 34-yr-old woman was referred to our hospital complaining of severe lumbar pain and high-grade fever (39°C) for the previous 4 days. Her medical history disclosed that the patient suffered from hypothyroidism, β-thalassemia minor, and allergy to penicillin. Four weeks before, the patient had delivered her second child and had been given epidural analgesia for pain relief during labor. According to the patient’s medical records, an epidural catheter had been inserted at the L2-3 interspinous space using an 18-gauge Tuohy needle after several attempts. She had no symptoms of systemic infection, and there was no evidence of a local skin infection. The skin was prepared with povidone iodine 10% from a multi-use bottle (Betadine; Asta Medica S.A., Madrid, Spain), which was wiped off vigorously from the puncture site with sterile sponges. A sterile technique was used, which included the use of a face mask, cap, gown, and sterile gloves. After performing a test with lidocaine and epinephrine that was negative, an infusion of 0.0625% bupivacaine and 2 μg/mL of fentanyl was administered at 12 mL/h. A bacterial filter attached to the epidural catheter was used during all the infusion. The catheter was left in place until delivery 7 h later. There were no untoward obstetric events. At admission, standard postero-anterior and lateral lumbar spine radiographs were taken and showed signs of spondylodiscitis at the L1-2 intervertebral space. Magnetic resonance imaging (MRI) revealed changes in the intervertebral disk, together with erosions of the adjacent vertebral endplates at the L1-2 interspace (Fig. 1). Laboratory findings included a white blood count of 6400 with 61% neutrophils, an erythrocyte sedimentation rate of 50 mm/h, and a serum C-reactive protein level of 80 mg/L. Diabetes mellitus, urinary tract infection, and human immunodeficiency virus infection were eliminated by the appropriate laboratory tests. A fine-needle biopsy guided by computed tomography was performed. Culture of the removed tissue in anaerobic conditions grew a P. acnes strain, which was sensitive to β-lactams, clindamycin, vancomycin, and ciprofloxacin. The patient was treated with ciprofloxacin for 6 wk and a lumbosacral corset. Three weeks after starting the treatment, she became asymptomatic and her C-reactive protein level and erythrocyte sedimentation rate became normal.

Figure 1: Magnetic resonance imaging (MRI) of the lumbar spine demonstrating signal changes at the margins of the L1-2 vertebrae and associated disk destruction.

Discussion

No case of spondylodiscitis by P. acnes after epidural analgesia for labor and delivery has been reported. P. acnes is a nonspore, gram-positive, anaerobic bacterium of the diphtheroid family. It is a normal saprophyte in the skin, sebaceous glands, and hair follicles. P. acnes has been documented as the causative agent in spontaneous discitis in some immunocompetent individuals ^(3–5). P. acnes discitis has also been reported after spinal and epidural anesthesia by local contamination ^(6,7). Early diagnosis of spondylodiscitis is often difficult. Careful physical examination, neuroimaging studies, and a high degree of suspicion are required to arrive at a correct diagnosis. Patients often present with increasing lower-back pain, tenderness, and restricted motion of the lumbar spine together with fever, and blood tests indicating infection ^(2–7). The interval between the inoculation of P. acnes and the appearance of clinical signs of infection varies from two to six weeks. However, P. acnes may reside intracellularly and remain in a dormant state for months, resulting in a longer incubation period ⁽⁸⁾. Diagnosis of spondylodiscitis is confirmed by MRI, which shows early inflammatory changes of disk, bone, and surrounding soft tissues. In addition, MRI is the most sensitive and specific imaging technique for the diagnosis of spinal infection ⁽⁹⁾.

Clinical suspicion and imaging studies may signal the correct diagnosis, but tissue sampling is often required for characterization of the infecting organism. Occasionally, microbiologic diagnosis can be made from peripheral blood cultures ⁽¹⁰⁾. When blood cultures are nonconfirmatory, computed tomography-guided fine-needle aspiration can provide an adequate sample for microbiological study in 18% to 86% of cases ^(1,10). However, if a precise diagnosis remains elusive, biopsies by open or laparoscopic surgery are alternative methods for bacteriologic verification ^(6,11). The current management for spondylodiscitis includes a six- to eight-week course of specific IV antibiotic treatment as well as bedrest followed by lumbosacral immobilization with a corset for pain relief. Surgical treatment, with debridement and bone grafting, is generally reserved for cases with marked spinal instability or for those with neurological involvement ⁽⁶⁾.

The mechanism of infection in our patient was probably by direct contamination. No other sources of infection were found, and the patient was neither debilitated nor immunocompromised. Full aseptic precautions were taken during catheter insertion and epidural analgesic delivery by continuous infusion through a closed system used in the ward. The reported antimicrobial activity of local anesthetics can be insufficient for decreasing the risks for infection related to epidural cannulation ⁽¹²⁾. The most probable explanation for the occurrence of discitis in our patient rests upon the difficulties encountered with epidural needle insertion because several puncture attempts were required. In our opinion, these technical difficulties might have facilitated contamination, or they might have favored epidural hematoma formation that, in turn, would have acted as a nidus for infection ⁽¹³⁾. However, spondylodiscitis may occur after epidural catheterization without the coexistence of a known causal event ⁽¹⁴⁾. Several features point toward a strong causal relation between the epidural analgesia and the spondylodiscitis in our patient such as length of the incubation period, location of the lesion, puncture level, and growth of an organism of cutaneous origin. This observation should alert the anesthesiologist to the fact that infection, although rare, may even occur after an uneventful epidural analgesia in the healthy parturient and reinforces the importance of performing the procedure under a strict aseptic technique.