Radiation proctitis-related lumbar spondylodiscitis due to Actinomyces odontolyticus: a rare finding

Abstract

Lumbar spondylodiscitis due to radiation proctitis-related fistula is a rare finding in the literature. After having isolated Actinomyces odontolyticus, a rare finding in the osteomuscular system, we present one of such cases.

A 75-year-old patient with a history of rectum adenocarcinoma, submitted to surgery and radiotherapy, presented himself in our emergency department with a 3-month history of lumbar pain radiating to both legs. Physical examination was compatible with cauda equina syndrome and subsequent investigation revealed L4–L5 spondylodiscitis. Despite a 6-month antibiotic therapy regimen, the symptoms recurred. Intravertebral disc biopsy revealed A. odontolyticus and directed antibiotic therapy was started. However, the symptoms recurred after a new 6-month antibiotic therapy regimen, this time with rectal purulent drainage. Additional study revealed two rectal fistulae. It was assumed those were caused by radiation proctitis and constituted the primary cause of spondylodiscitis. Laminectomy was performed with a satisfactory clinical response.

Background

Infection secondary to fistulas is in itself a rare occurrence. The more typical finding is fistula formation secondary to an infectious event. Pelvic fistulas are mainly due to infection, inflammatory bowel disease and iatrogenia (for example, surgery or radiotherapy), usually in the context of prostate and colonic cancer.^{1 2} Iatrogenic complications related to radiotherapy are most commonly pelvic inflammation, ileus and oedema, with diarrhoea, intestinal perforation, fistula, cystitis and sacral osteolysis being much less frequent (4.9% of patients in each case).³ Pelvic fistulas can involve the spine, causing spondylodiscitis and other complications, such as epidural abscesses, which can cause cauda equina syndrome. However, these are not common complications of pelvic fistula. To the best of our best knowledge, this is the third case of a bone-related complication literature described, with the other ones being related to Crohn’s disease and surgical procedures.^{1 4 5}

Case presentation

A 75-year-old male patient presented himself to our emergency department in December 2017 with a 3-month history of lumbar pain radiating to both legs. Any other symptoms were denied, mainly oral or dental disease, fever, palpitations, syncope, chest pain, cough or gastrointestinal symptoms. He had a history of rectum adenocarcinoma (T3N0MO) and was operated on in 2017 and submitted to radiotherapy. Because of an anastomosis leak, he was again submitted to surgery, this time for a Hartmann procedure. Physical examination showed no abnormalities except slight pain on palpation of the lumbar column, rectal purulent exudate, and reduction of bilateral proximal strength on neurological examination.

Investigations

Laboratory findings showed normocytic hypochromic anaemia (83 g/L), leucocytosis (white blood count 16 x 10⁹ L) with neutrophils and C reactive protein elevation (CRP; 22.53 mg/dL). No other analytical abnormalities were present. Dorsal and lumbar radiography had no alterations indicating any specific disease. However, CT scan revealed findings compatible with spondylodiscitis.

Differential diagnosis

Because of the clinical, laboratory and radiological findings, spondylodiscitis was suspected. As such, a MRI of the spinal column was performed. This revealed characteristics compatible with spondylodiscitis in L5–S1 and peripheral oedema. An epidural septated intracanal collection was also observed at the level of L5–S1, as well as signs of infectious epiduritis and arachnoiditis at the level of L5–S2 (figure 1). Together with the neurological deficits presented by the patient, these findings established the diagnosis of cauda equine syndrome.

Figure 1.

First MRI. Epidural septated intracanalar collection at the level of L5–S1, as well as signs of infectious epiduritis and arachnoiditis at the level of L5–S2.

Rectosigmoidoscopy was performed and within the site of the Hartman procedure suture, two orifices were observed, suggesting fistulas. No masses suggesting neoplasm were observed.

Treatment

Empirical antibiotic therapy was started with vancomycin, covering methicillin-resistant Staphylococcus aureus and Enterococcus (completing 7 days) and ceftriaxone covering Enterobacteriaceae (completing 14 days). Blood cultures, rectal exudate and urine cultures were all negative. Echocardiogram did not show any abnormalities suggestive of concomitant endocarditis. As the patient revealed a good therapeutic response, amoxicillin and trimethoprim sulfamethoxazole were started and continued for a 6-month period with clinical and radiological improvement.

Outcome and follow-up

After a total 6 months of antibiotic therapy and 1 month afterwards, with clinical and laboratory improvement (leucocytes 7400 cells, with 5180 neutrophils, CRP <0.5 mg/dL and erythrocyte sedimentation rate (ESR) of 18 mm at 4 months of antibiotic therapy), the patient reinitiated lumbar pain, this time accompanied by paraesthesias and proximal leg muscle strength reduction. He had no other symptoms. Recurrence of spondylodiscitis was suspected, confirmed by CT scan (figure 2), and the patient was admitted to the hospital. As previous cultures had been inconclusive as to the causative microorganism, intervertebral disc biopsy was performed. Actinomyces odontolyticus was isolated and ampicillin was started. After 13 days of antibiotic therapy, improvement in symptom was observed, allowing patient discharge. Laboratory findings at this time revealed leucocytosis (9x10⁹ leucocytes, with 7230 neutrophils), CRP of 1.13 mg/dL and ESR of 36 mm. At this point, amoxicillin was started, with the intention of completing a 1-year antibiotic therapy regimen.

Figure 2.

CT scan showing recurrence of spondylodiscitis.

However, 6 months after discharge, new onset pain worsening and anal purulent exudate was observed. No elevation in leucocyte value was observed, but ESR and CRP values had augmented (70 mm and 4 mg/dL, respectively). MRI showed the presence of spinal compression from L4 to S2, as well as a purulent collection within the intervertebral disc from L5 to S1. Posterior epidural lipomatosis from L3 to S1 was also observed (figure 3). Rectosigmoidoscopy was performed, but no therapy was possible. Purulent exudate isolated a Pseudomonas aeruginosa sensitive to ceftazidime and this antibiotic was started, completing 14 days, continued by ciprofloxacin after discharge. With this regimen, the patient showed some symptomatic improvement. However, laminectomy was considered as the best option, and it was performed 1 month later, with resolution of the paraesthesias and significant functional improvement. Fistula drainage was tried two times but failed, the first time because the site of the orifices was inaccessible and the second time because the orifices were not observed. At the present time, the patient is waiting for the surgical resolution of these fistulas, with symptomatic control and no further need for hospital admissions.

Figure 3.

Second MRI. Spinal compression from L4 to S2 and a purulent collection within the intervertebral disc from L5 to S1.

Discussion

Due to increased sensitivity and specificity of imaging methods, the diagnosis of spondylodiscitis has increased over the last 10 years.⁶ Nowadays, it represents 2%–7% of all musculoskeletal infections, with mortality ranging between 2% and 20% in developed countries.⁶ It usually affects men more than women in a ratio male-female between 2:1 and 5:1.⁶ It has two incidence peaks, one in the second decade of life and the other one between the sixth and eighth.⁶ Risk factors include spinal surgery, obesity, diabetes mellitus, substance abuse, chronic infection, immunosuppression, or incompetence and poor nutritional status.⁶ In patients with spondylodiscitis, the primary source of infection is the genitourinary system, followed by the respiratory tract.⁵ Clinically, often with a non-specific presentation, the most common symptom is back pain (85%), followed by fever (48%) and paresis (32%).⁶ S. aureus followed by Gram-negative bacteria are the most common organisms isolated in developed countries. Additionally, it is also the most common microorganism related to epidural abscesses.⁶ Worldwide the most common organism is Mycobacterium tuberculosis. Other common bacteria include Brucella (2%–60% of all brucellosis cases, with 60% of these infecting the lumbar column),⁷ Escherichia coli and, to a lesser extent, Pseudomonas spp.⁸ Less common microorganisms include fungi and parasites.⁶ Several routes of infection spread have been identified, which include haematogenous, spreading from contiguous tissues (most common route, and in this case, a bone-enteric fistulae) and direct inoculation.

The most sensitive laboratory biomarker is CRP, being elevated in up to 90% of spondylodiscitis cases.⁶ Leucocytosis can be absent and ESR is less specific and can be increased in up to 50% of patients who have good therapeutic response.⁶ Procalcitonin, although useful in most cases to distinguish bacterial infection from inflammation, shows less sensitivity than CRP in these cases.⁹ Blood culture is positive in only 60% of patients, and the most useful diagnostic method for microorganism isolation is tissue biopsy, with a sensibility of up to 90%. This was also shown in our patient, as haemocultures, purulent exudative and urine cultures were all negative, the only microbiological isolation being through tissue biopsy.⁶

The most useful imaging method for the diagnosis of spondylodiscitis is MRI, having a sensitivity and specificity of 96% and 94%, respectively.⁶ Its usefulness is increased by the fact that it can provide detailed information of contiguous tissues, such as the epidural space. It can also be used in the follow-up, but it should not be performed within the first 6 weeks of treatment, as radiological signs can show a false worsening of the lesions.¹⁰ Another radiological approach in follow-up is through positron emission tomography scan, as normal bone tissue does not uptake radionuclides. However, because it does not distinguish between conditions as neoplasms and osteochondritis, it is not useful at the present time to diagnose osteomyelitis. CT scan has less sensitivity and specificity than MRI; however, it is useful to guide tissue biopsy, and is used in the context of our patient.

Spondylodiscitis treatment involves at least 6 weeks of intravenous antibiotic therapy, with good clinical, laboratory and radiological results if the diagnosis is precocious. However, there are some clear situations where surgery should be the primary treatment, namely the worsening of neurological deficits, signs of spinal cord compression and spinal instability. In all other situations, antibiotic therapy can be tried, and surgery is only indicated if medical therapy fails.^{11 12} In our case, we decided on a 6-month course due to the microorganism that was isolated, A. odontolyticus. After directed antibiotic therapy proved inefficient, laminectomy, which is, together with spinal fusion, one of the possible procedures, was performed, with good results.¹⁰

Actinomycosis is a rare, chronic granulomatous disease caused by the Actinomycetaceae family of microorganisms, with A. israelli being the most common pathogen. Actinomyces are Gram-positive, slow-growing, anaerobic or aerobic facultative rods. They are commensal to the gastrointestinal and urological tracts and infect middle-aged individuals, more often between the ages of 20 and 60 years with the male gender being affected three times as much as the female gender.^{13 14} Risk factors include immunosuppression (for example: diabetes, immunosuppressive drugs, infection with the HIV) and male gender. Infections can lead to necrosis, abscess formation and fistulas. In half of the reported cases, it affects the cervical area, with other sites the like abdomen or the thorax (20% each) being rarer. Definitive diagnosis is often only possible with microorganism isolation.¹⁴ However, it is estimated that only 50% of microorganisms are isolated due to a panoply of factors. Pus is the most appropriate sample to isolate Actinomyces, and it was from pus that Actinomyces was isolated in our case. The microorganism is susceptible to penicillins. Classically, it requires treatment for a duration of 6 months to 1 year,¹⁵ but more recent regimens involve severity of disease and infection site in their algorithms. In fact, in our case, even with antimicrobial regimens of 6 months, we were not successful in treating the patient, most likely due to the site of infection and probable consequential low antimicrobial therapy penetration, despite the use of appropriate antibiotics. The prognosis depends on the site of infection, with a mortality ranging between 0% and 28%, being worse if there is central nervous system involvement.¹⁵

Musculoskeletal actinomycosis, especially bone infection, is a rare finding, and even rarer if central nervous system focalisation is also present. Epidural abscesses count only as 6% of central system actinomycosis. Central system actinomycosis is usually due to infected adjacent tissues.¹⁵ This was probably the mechanism in our case, as the patient had been subjected to surgery and radiotherapy due to neoplasia, both being risk factors for abdominal Actinomyces infection.¹⁴ Clinical presentation of bone infection with central nervous system focalisation is usually in the form of cauda equina syndrome that can have an insidious or acute onset,^16–18 and this was the case in our patient.

Patients in whom high bioavailability can be achieved, oral antibiotic therapy is an option.^{19 20} Treatment failure factors include age above 50 years, neurological compromise and comorbidities. Surgical intervention should be considered early in the natural history of the disease followed by an extended course of pathogen-directed therapy in similar refractory and complicated cases for definite source control of vertebral Actinomyces infection.

Learning points.

• Antibiotic therapy should be only initiated after tissue aspiration biopsy, which should be performed in all cases. An exception should be made in septic patients, in whom the delay in initiation of antibiotic therapy can result in death. The duration of antibiotic therapy should not be inferior to 6 weeks of endovenous or high bioavailability oral antibiotic. Surgery should be the first-line treatment modality in patients with worsening neurological signs, spinal cord compression and spinal instability. It should be also performed in patients who are revealed to be refractory to antibiotic therapy.

• Clinical and laboratory follow-up should be maintained at least for the duration of antibiotic therapy to detect early signs of recurrence, such as new onset or worsening pain and C reactive protein elevation. After treatment suspension, MRI and CT scan can be useful to distinguish between disease recurrence and chronic pain, a possible/common sequela. Radiological surveillance is not as useful in the first weeks as it is later in the course of the disease. Follow-up should be maintained at least until there is complete clinical and analytical stabilisation, which should then be confirmed by radiological proof of non-active disease, ideally through MRI.

Ethics statements

Patient consent for publication

Obtained.