- Email: [email protected]
Spontaneous pyogenic vertebral osteomyelitis in nondrug users
Spontaneous Pyogenic Vertebral Osteomyelitis in Nondrug Users Joan M. Nolla, Javier Ariza, Carmen Go´mez-Vaquero, Jordi Fiter, Joaquı´n Bermejo, Josep Valverde, Daniel Roig Escofet, and Francesc Gudiol Objective: To analyze the clinical, microbiological, and radiologic features of patients without drug addiction suffering from spontaneous pyogenic vertebral osteomyelitis. Methods: We collected all microbiologically proved cases of pyogenic vertebral osteomyelitis seen between January 1980 and December 1999 in a teaching hospital. Patients with prior spinal instrumentation or surgery and injection drug users were excluded. Results: Sixty-four patients, with a mean age of 59 ⴞ 17 years, were identified. In 29 (45%) patients, 1 or more underlying medical illnesses were found. The mean duration of symptoms before hospital admission was 48 ⴞ 40 days. Neurologic impairment was present in 18 (28%) patients. Staphylococcus aureus and gramnegative bacilli, mainly Escherichia coli, were the predominant etiologic agents. Blood cultures were positive in 72% (46/64) of cases. The cultures of spinal specimens obtained by x-ray– guided biopsy were positive in 52% (11/21) of cases, and those obtained by open biopsy in 75% (15/20) of cases. Plain radiography showed abnormalities in all but 7 patients. Fifty-one of 53 (96%) technetium Tc 99m diphosphonate bone scans and 40 of 44 (91%) gallium citrate Ga 67 bone scans showed increased uptake of tracers in the involved area. Paraspinal and epidural extension was found on computed tomography and/or magnetic resonance imaging in 74% (39/53) of cases; the presence of an extra-vertebral extension was not associated with the development of neurologic findings in many patients. Two patients died in relation to the infectious process and 3 relapsed; functional sequelae often were found. Conclusions: Spontaneous pyogenic vertebral osteomyelitis in nondrug users is a disease that affects mainly older patients suffering underlying medical illnesses. S aureus and E coli are the main causative microorganisms. Positive blood cultures frequently aided the diagnosis. Extra-vertebral extension is frequent but does not indicate a worse prognosis. Although life outcome is good, functional sequelae are common. Diagnostic delay before admission is a concern, and the physician should be alert to the possibility of this condition in patients with back or neck pain. Semin Arthritis Rheum 31:271-278. Copyright 2002, Elsevier Science (USA). All rights reserved. INDEX WORDS: Osteomyelitis; vertebral osteomyelitis; spondylodiskitis; spondylitis; spinal infection; spine.
From the Rheumatology Department and Infectious Diseases Department, Ciutat Sanita`ria i Universita`ria de Bellvitge, Barcelona, Spain. Joan M. Nolla, MD, PhD: Professor of Medicine, Staff member, Rheumatology Department, Javier Ariza, MD, PhD: Professor of Medicine, Staff member, Infectious Diseases Department, Carmen Go´mez-Vaquero, MD, PhD: Staff member, Rheumatology Department, Jordi Fiter, MD: Staff member, Rheumatology Department, Joaquı´n Bermejo, MD, Fellow, Infectious Diseases Department, Josep Valverde, MD, PhD: Professor of Medicine, Staff member, Rheumatology Department, Daniel Roig Escofet, MD, PhD, Pro-
fessor of Medicine, Chair of Rheumatology Department, Francesc Gudiol, MD, PhD, Professor of Medicine, Chair of Infectious Diseases Department; Ciutat Sanita`ria i Universita`ria de Bellvitge, Barcelona, Spain. Address reprint requests to Dr. Joan M. Nolla, Rheumatology Service (Pl 10-2), Ciutat Sanita`ria i Universita`ria de Bellvitge, Feixa Llarga s/n, 08907 L⬘Hospitalet, Barcelona, Spain. E-mail address: [email protected] Copyright 2002, Elsevier Science (USA). All rights reserved. 0049-0172/02/3104-1048$35.00/0 doi:10.1053/sarh.2002.29492
Seminars in Arthritis and Rheumatism, Vol 31, No 4 (February), 2002: pp 271-278
271
272
NOLLA ET AL
P
YOGENIC VERTEBRAL osteomyelitis (PVO) has been regarded as an uncommon entity (1). Several investigators, however, have stressed that its incidence seems to be increasing as a result of the higher life expectancy of older patients with chronic debilitating diseases, the rise in the prevalence of intravenous drug abuse, and the increase in spinal instrumentations and surgery. A clear distinction should be made between postoperative and spontaneous PVO. Postoperative PVO is produced as the result of direct inoculation (2,3), whereas spontaneous PVO (3,4) is thought to be caused by hematogenous spread into the disc periphery or metaphysis of the vertebral bodies. The 2 forms have important clinical and microbiological differences and require different approaches. In spontaneous PVO (4), the clinical picture and relative prevalence of causative microorganisms are substantially different in drug and nondrug users. In drug users, the disease affects mainly young patients, follows a septic course, and Staphylococcus aureus is the etiologic agent in the great majority of cases; in these individuals, the presence of a strong risk factor makes the diagnosis easy. In nondrug users, PVO occurs predominantly in aged patients, septic manifestations are not always present, and a wide variety of gram-positive and gram-negative organisms act as causative agents; in this population, the diagnosis is easily missed. Until recently, the information available on PVO was limited (5). In the last 10 years, several series have been published (6-22); however, some were small (8,9,13,14) and others focused on 1 particular etiologic agent (19-21) or radiologic technique (8,18,22). As a result, certain diagnostic and therapeutic issues relating to the practical management of patients remain to be clarified. In this study, we review our experience with spontaneous PVO in nondrug users from the dual perspective of rheumatic and infectious disease specialists. We focus on decisive clinical and microbiological aspects of the disease and evaluate the contributions of modern radiology to the diagnostic and therapeutic management of patients. PATIENTS AND METHODS
Patients The charts of all adult patients in whom vertebral osteomyelitis was diagnosed between January
1980 and December 1999 in the Ciutat Sanita`ria i Universita`ria de Bellvitge, a 1,000-bed tertiary teaching hospital in the Barcelona area, were reviewed. Patients with prior spinal instrumentation or surgery, injection drug users, and those with tuberculosis or brucellosis were excluded. We included only patients in whom a definitive diagnosis of PVO was made by the presence of 3 criteria: 1) compatible clinical picture (spinal pain and localized tenderness or limited range of motion), 2) compatible imaging findings (decreased height of the intervertebral disk with or without erosions of the end plates on adjacent vertebral bodies on plain radiographs and/or increased uptake of tracers on technetium Tc 99m diphosphonate or gallium citrate Ga 67 bone scans), and 3) isolation of a pyogenic organism in blood cultures or spinal biopsy specimens. Sixty-four patients met the criteria; the cases were evenly distributed throughout the study period. A diagnosis of probable PVO was made in 11 patients (6 men and 5 women; mean age, 69 years) during the same period. They had a compatible clinical picture, suggestive imaging findings, and an appropriate response to antibiotic therapy, but no organism was isolated. Furthermore, 28 patients with tuberculous vertebral osteomyelitis and 35 patients with brucellar vertebral osteomyelitis were identified. Data Collection Every patient’s chart was retranscribed onto an standard itemized form and included the following information: 1. 2. 3. 4.
Demographic data Underlying medical illnesses Vertebral level involved Clinical presentation a. Duration of symptoms before diagnosis b. Type of spinal pain: mechanical (pain while moving) or inflammatory (pain at rest) c. Findings on spinal physical examination (tenderness, limited range of motion) d. Neurologic findings e. Maximal recorded temperature (°C) prior to diagnosis 5. Bacteriologic diagnosis 6. Presumed sources of infection and concomitant infectious processes
PYOGENIC VERTEBRAL OSTEOMYELITIS
7. Laboratory tests a. Erythrocyte sedimentation rate (ESR) evaluated by Westergren method b. White blood cell (WBC) count 8. Radiologic tests a. Plain radiography b. Technetium bone scan c. Gallium bone scan d. Computed tomography (CT) e. Magnetic resonance imaging (MRI) 9. Treatment and outcome Bacteriologic Studies A minimum of 2 blood samples were obtained from all patients before the start of antibiotic therapy. Spinal specimens were cultured in 5% horse blood agar, chocolate agar, and MacConkey agar plates and in thioglycolate medium. Microorganisms were identified according to standard criteria (23). Statistical Study The relation among categorical variables was established by the chi-square test. Correlations were made by means of Spearman’s test. Differences among groups of patients were assessed by analysis of variance. A P value of ⬍ .05 was considered significant. RESULTS
Demographic Data Of the 64 patients, 36 (56%) were men and 28 (44%) were women, with ages ranging from 18 to 84 years (mean, 59 ⫾ 17 years). Thirty-eight (59%) patients were over 60 years, and 7 (11%) were under 25. Underlying Medical Illnesses In 29 (45%) patients, 1 or more underlying medical illnesses were found: diabetes mellitus in 18, liver cirrhosis in 6, renal failure in 5, malignancies in 2, rheumatoid arthritis treated with glucocorticoids in 1, and congestive heart failure in 1. These patients were on average older than the 35 patients without relevant medical illnesses (65 ⫾ 8 years v 55 ⫾ 20 years; P ⬍ .05). Vertebral Level Involved Sixty-one (95%) patients had involvement of only 1 motion segment (2 adjacent vertebrae and
273
the intervening disc): the lumbar area in 40 cases, the thoracic area in 17, and the cervical area in 4. Three (5%) patients had involvement of several segments, which were contiguous in 2 cases (1 in lumbar area and 1 in thoracic area) and not contiguous in 1 (lumbar and cervical areas). Clinical Presentation The mean duration of symptoms before hospital admission was 48 days ⫾ 40 days. No significant correlation between the age of the patient and the duration of symptoms was found. No significant difference was found in the duration of symptoms among the etiologic agents. Back or neck pain was reported in all cases; it was inflammatory in 48 (75%) patients and mechanical in 16 (25%). Physical examination revealed localized spinal tenderness in 62 (97%) patients and a limited range of motion in 60 (94%); all patients had at least 1 of these findings. Neurologic impairment was present in 18 (28%) patients: 11 (17%) had radiculopathy, 5 (8%) had lower limb weakness, and 2 (3%) with cervical involvement due to S aureus had quadriparesis. Thirteen (20%) patients had an axillary temperature ⬍37°C. In 11 (17%) patients, the temperature ranged from 37°C to 38°C, and it was ⬎38°C in the remaining 40 (63%) patients. Bacteriologic Diagnosis In all cases, the infection was monomicrobial; the organisms isolated are listed in Table 1. S aureus and E coli were the predominant etiologic agents. S aureus was isolated in 26% (10/38) of patients over 60 years, in 50% (13/26) of those under 60 years, and in 100% (7/7) of those under 25 years. Seventy-three percent (11/15) of patients with PVO due to E coli were over 60 years of age. Table 2 shows the diagnostic procedures used. Etiologic diagnosis was established on the basis of blood cultures alone in 39 cases and on the basis of biopsy specimens alone in 18 cases (S aureus in 6 patients, Streptococcus agalactiae in 3, Pseudomonas aeruginosa in 2, and Streptococcus salivarius, Streptococcus mitis, Streptococcus sanguis, Haemophylus paraprophylus, Proteus vulgaris, Proteus mirabilis and Eikenella corrodens in 1 patient each). We identified the same organism in the 7 cases in which blood and spinal cultures gave positive results. Blood cultures were positive in 46 of 64 (72%)
274
NOLLA ET AL
Table 1: Organisms Isolated From 64 Patients With Pyogenic Vertebral Osteomyelitis
Bacteria
No. (%) of Cases
Staphylococci Staphylococcus aureus* Staphylococcus epidermidis Streptococci Streptococcus sanguis Streptococcus agalactiae Streptococcus salivarius Streptococcus mitis Streptococcus bovis Streptococcus pneumoniae Gram-negative bacilli Escherichia coli Pseudomonas aeruginosa Eikenella corrodens Proteus mirabilis Proteus vulgaris Haemophylus aprophylus Haemophylus paraprophylus Other organisms Enterococcus faecalis Bacteroides fragilis
25 (39) 23 2 12 (19) 5 3 1 1 1 1 25 (39) 15 3 2 2 1 1 1 2 (3) 1 1
*One case, methicillin-resistant (MRSA).
cases: 34 of the 40 (85%) patients with axillary temperature ⬎38°C, 6 of the 11 (54%) patients with axillary temperature 37°C to 38°C, and 6 of the 13 (46%) with axillary temperature ⬍37°C (2 of whom had endocarditis). Blood cultures were positive in 19 of 25 (76%) patients with staphylococcal infection or gram-negative bacterial infection and in 6 of 12 (50%) patients with streptococcal infection. Biopsy was performed in 36 patients. The specimen was obtained by percutaneous x-ray (usually CT) guided biopsy in 16 cases and by open biopsy in 15. In 5 patients, both biopsy techniques were performed. The yield of x-ray-guided biopsy was 52% (11/21) and of open biopsy was 75% (15/20). In 3 cases, in which needle biopsy and/or open biopsy were performed after antibiotic therapy had been initiated, the cultures of the obtained specimens were negative. Among the 5 cases with percutaneous and open biopsy, 1 gave positive results by both techniques; the cultures of the other 4 spinal specimens obtained by x-ray-guided bi-
opsy were negative, and 2 cultures obtained by open biopsy were positive. No complications related to biopsy were documented. Presumed Sources of Infection and Concomitant Infectious Processes PVO was community-acquired in all patients except for 1 hemodyalized older man, in whom the condition was secondary to a cannula-related sepsis due to methicillin-resistant S aureus. Overall, a source of infection or a concomitant infectious process was identified in 30 (47%) of patients. Lower urinary tract infection was the presumed source of infection in 13 patients, who developed compatible symptoms before the onset of back pain. E coli was the causative organism in 10 cases, P aeruginosa in 2, and P mirabilis in 1. In 2 patients, the urinary infection was proven microbiologically at the time of PVO diagnosis. Two patients with PVO due to S aureus had a cutaneous ulceration which was considered the source of infection. In 1 patient with PVO due to E coli, a colonic endoscopy with biopsy of a polyp had been performed 2 weeks before the onset of symptoms. One patient with Bacteroides fragilis infection had a periapendicular abscess. In the patient with Streptococcus bovis infection, colonic diverticulosis was diagnosed after the diagnosis of PVO. In 1 patient with infection due to E corrodens, a rectal neoplasm was diagnosed during the hospital stay. Two patients with cirrhosis had spontaneous peritonitis due to E coli. In 1 cirrhotic patient with PVO due to S epidermidis, an infection of a LeVeen shunt was identified. Blood cultures were positive in all cirrhotic patients.
Table 2: Diagnostic Procedures Used in 64 Cases With Pyogenic Vertebral Osteomyelitis
Diagnostic Procedures Blood cultures Blood cultures and x-ray–guided biopsy Blood cultures and open biopsy Blood cultures, x-ray–guided biopsy, and open biopsy
No. (%) of Patients 28 (44) 16 (25) 15 (23) 5 (8)
PYOGENIC VERTEBRAL OSTEOMYELITIS
Three patients with PVO due to S aureus presented other concomitant staphylococcal osteoarticular infections: knee arthritis, sacroiliitis, and elbow bursitis. Four patients suffered from endocarditis: S sanguis was the causative organism in 2 cases and E corrodens and S epidermidis in 1 each. Laboratory Tests At admission, the mean ESR (⫾ SE) was 77 mm/h ⫾ 33 mm/h (range, 8-140 mm/h); 6 (9%) patients had an ESR below 30 mm/h. The mean WBC count (⫾ SE) was 9.7 ⫾ 6 ⫻ 109/L (range, 1.7-37 ⫻ 109 /L); the WBC count was above 11 ⫻ 109/L in 23 (36%) patients. Radiologic Tests All patients underwent plain radiography of the spine, 53 (83%) had technetium bone scans, 44 (69%) had gallium scans, 36 (56%) had CT, and 33 (52%) had MRI. CT and MRI were performed only after plain radiography and bone scanning. CT and/or MRI were available in all cases with neurologic findings. Plain radiography revealed abnormalities in 57 of 64 patients. Increased uptake on technetium and/or gallium bone scans was observed in the 7 patients with negative x-rays; these patients had symptoms of short duration (mean, 14 days; range, 7-21 days). Fifty-one of 53 (96%) technetium and 41 of 44 (91%) gallium bone scans showed increased uptake of tracer in the affected vertebral level. In the 2 patients with negative technetium scanning and whose duration of disease was less than 15 days, gallium scans were positive. Among the 4 patients with negative gallium bone scans, all of whom presented pathologic findings on plain radiography, technetium scans were positive in 3 and scanning was not performed in 1. In 53 patients, a CT and/or an MRI was performed (CT alone in 20 cases, MRI alone in 17, and both in 16), and changes consistent with PVO were demonstrated in all cases. In 15 patients in whom a gadollinium-enhanced MRI was performed, contrast uptake was observed. Paraspinal or epidural extension was seen in 39 (74%) (paraspinal alone in 16 cases, epidural alone in 5 cases, and both 18). Compression of the spinal cord was identified in 7 cases, but no change in spinal cord signal was found. CT and MRI data
275
were the same. Diagnostic delay and etiologic agents were similar in patients with or without extra-vertebral extension. Fifteen of the 39 patients with extension had neurologic findings: 9 patients had radiculopathy (7 with paraspinal and 2 with both paraspinal and epidural extension), 5 patients had lower limb weakness (1 with paraspinal and 4 with both paraspinal and epidural extension), and 1 patient, with extensions into both areas, had quadriparesis. In 1 patient with quadriparesis, only signs of subluxation were observed on CT. Treatment and Outcome Four patients died. Two of them, with S aureus bacteremia, died of infection on days 4 and 8 after beginning appropiate antibiotic therapy. In the other 2 cases, death was not infection related but attributable to traumatic splenic rupture in 1 case on day 15 and myocardial infarction in the other on day 62. Fifty-seven patients received intravenous antibiotics during 4 to 6 weeks; 20 of them were given further oral therapy for 2 to 4 weeks. Four patients received only oral treatment for 8 weeks; 3 with PVO due to E coli were given fluoroquinolones and 1 with S aureus infection received a combination of rifampicin and ciprofloxacin. No antibioticrelated complications were noted. Bed rest was prescribed in all cases with external immobilization when movement was authorized. Surgery was indicated in 4 patients with severe neurologic findings. However, patients who had an open biopsy usually were drained during the surgery. The mean hospital stay was 57 ⫾ 37 days, and no differences were observed between patients with or without underlying diseases (65 ⫾ 47 days v 50 ⫾ 24 days; P ⫽ ns). Information regarding back pain at discharge was obtained in 50 cases: 16 patients had no pain, 30 had mechanical pain, and 4 had inflammatory pain. All patients with radiculopathy recovered. The 2 patients with quadriparesis had lower limb weakness, and 3 of the 4 surviving patients with lower limb weakness recovered; 1 had persistent symptoms. Forty-one patients were seen regularly in our outpatient clinics after discharge (mean follow-up, 68 ⫾ 50 weeks). Three patients who had been treated with intravenous antibiotics for 6 weeks relapsed. Two had diabetes with PVO due to
276
NOLLA ET AL
S aureus, and 1 patient with cirrhosis had P aeruginosa infection. In patients with staphylococcal infection, extravertebral extension had been identified by CT and/or MRI (paraspinal extension in 1 case and both paraspinal and epidural extension in the other); the patient with epidural extension had undergone surgical drainage. Relapses occurred at 6 and 5 months, respectively, after the end of antibiotic therapy. The patient with cirrhosis and P aeruginosa infection showed no extra-vertebral extension on CT. Relapse occurred 3 months after the end of antibiotic therapy. DISCUSSION
We present a large descriptive series of patients with vertebral osteomyelitis seen in our hospital for the last 20 years. Patients with prior spinal instrumentation or surgery and injection drug users were excluded, since both these population groups present specific differential characteristics. In addition, we excluded patients with brucellar or tuberculous vertebral osteomyelitis, which in our and in other Spanish investigators’ (24) experience accounts for 50% of cases. Furthermore, we analyzed only microbiologically proven cases, which represented 85% of our patients with supposed spontaneous PVO. All these aspects should be considered when comparing our results with those of other more heterogeneous series (6-12,15-17). The demographic characteristics of our sample were similar to those reported elsewhere, in which spontaneous PVO was found to affect mainly older men (6,8,10,13,14,16,17), who suffered from underlying medical illnesses in nearly half the cases (6,12,14,16,17). The medical conditions most frequently associated with PVO in our patients, as in those of other series (6,12,14,16,17,21) were liver cirrhosis, chronic renal failure, and primarily diabetes mellitus. The delay in diagnosing PVO in our series and in previous reports (6,12,13,17) is of particular concern. Patients with apyrexia or low-grade fever who have mechanical-type back pain are often misdiagnosed. But physical examination provides an important aid to diagnosis, since all of our patients had spinal tenderness or limited range of spinal motion, 2 findings closely related to PVO (3). Once a clinical diagnosis of vertebral osteomyelitis has been established, an attempt to identify
the causative organism should be made (1,3,4). Blood cultures provide microbiological diagnosis in many cases, since PVO is usually a monomicrobial disease and isolates from positive blood cultures can be assumed to be responsible for the process (4). No discrepancies between isolates from blood and spinal specimen cultures were found in the 7 patients in which both procedures provided positive results. The reported yield of blood cultures in PVO ranges from 42% to 82% (12,17), depending on the population included. In our series of only microbiologically proven cases, blood cultures provided a substantial yield, positive in 72% of all patients and in 85% of those with a temperature above 38°C. In some cases, bacteremia was detected before the onset of PVO and was associated with the primary source of infection. A notable finding was that 46% of our patients with apyrexia at admission had positive blood cultures at some time during the course of the disease. When blood cultures are negative, spinal specimens must be obtained by closed percutaneous or open surgical biopsy; needle biopsy is usually preferred because of safety considerations. In our study, cultures were positive in 52% and 75% of cases, respectively, figures similar to previous reports (6,10,12,14,15,17). The number of biopsies was relatively small, and only 5 patients had both biopsy techniques; thus, we cannot determine which type of biopsy was most sensitive. A large number of gram-positive and gramnegative organisms have been identified as causative agents in PVO (4). Overall, S aureus has been the most commonly isolated organism, accounting for more than 50% of reported cases. Gram-negative bacilli produce almost 30% of cases and streptococci around 10%, although these percentages vary depending on the population studied. Carragee (16) stressed the relevance of low-virulence organisms, usually considered contaminants, but other series have not corroborated this finding. Prospective studies are needed to determine whether these microorganisms are responsible for certain cases of PVO. Our results reveal certain clinical-bacteriological associations. S aureus accounted for a greater proportion of cases in young patients, whereas E coli was present more frequently in the aged. The proportion of streptococcal infections should be noted; these occurred mostly in patients with no
PYOGENIC VERTEBRAL OSTEOMYELITIS
recognized source of infection. The association of endocarditis and PVO, observed in 6% of our patients, has been reported in the literature infrequently. The possibility of concomitant endocarditis should be considered in patients with PVO, especially if Streptococcus spp is isolated from blood cultures. The diagnosis of PVO may require a number of imaging techniques (1,3). Conventional radiography is necessary at both presentation and follow-up (1). Plain radiographs show changes suggesting PVO in most patients at this time of clinical evaluation (6,8,13,17). Technetium diphosphonate and gallium citrate isotope scans (25) have high sensitivity but low specificity and may be useful in diagnosing early vertebral involvement in patients with normal plain radiography, as was the case in 7 of 64 (11%) of our cases. Additional information can be obtained when multiple noncontiguous vertebrae are involved, a rare circumstance in nondrug users (10,12-14). Technetium diphosphonate is probably the most suitable tracer because it is simpler and safer to use. CT is the technique of choice for the guidance of needle biopsies and provides useful information on bone changes, disc space narrowing, and soft tissue collections (26). MRI is an excellent technique for the diagnosis of vertebral osteomyelitis (1,27). Findings are seen early (28), and higher contrast resolution detects marrow infiltration and intradural disease (8,28) with sensitivity and specificity both above 90% (28). Unfortunately, MRI is expensive and not available in some centers. However, MRI should be performed if available. The information it provides, mainly on the presence of extra-vertebral extension, may be very useful if the clinical outcome during follow-up is unsatisfactory. In many cases, the diagnosis of PVO can be made on the basis of plain radiography, reserving radionuclide bone imaging techniques for initial evaluation of suspected cases with normal plain radiography and CT when needle biopsy is required. Paraspinal and/or epidural extension is a frequent finding when CT and MRI are used for the diagnosis of PVO (74% of the patients in our series). Like other authors (8), we used the term extension, since we did not attempt to differentiate diffuse inflammation from a true abscess, as that may be very difficult.
277
No relationship was found between extension, microorganism type, or other clinical findings. In fact, in most patients this radiologic finding had no clinical consequences. Twenty-eight percent of our patients with PVO had concomitant neurologic findings, a prevalence within the range previously reported (20% to 52%) (12,13). Radiculopathy was associated with paraspinal extension, and lower limb weakness or quadriparesis occurred with epidural extension. However, the presence of epidural extension was not necessarily associated with the development of neurologic findings. Medical treatment alone, including antibiotic therapy, bed rest, and immobilization with orthosis, is suitable for most cases of PVO; surgical intervention is reserved for the management of complications or for the failure of medical therapy (1,3). The main indications for surgery include open biopsy, spinal cord compression with significant neurologic deficit, large abscess, and substantial kyphotic deformity. The therapeutic approach should be based on the clinical manifestations and not on the imaging findings (29). The optimal duration of therapy is not defined, since nearly all reports of PVO are retrospective. Usually, antibiotics are given intravenously for at least 4 to 6 weeks, which is the main reason for the long hospital stay, and some clinicians use oral antibiotics for an additional several weeks (3). Recently, oral antibiotics as the only or main therapy has been increasingly recommended for different types of osteomyelitis and other serious osteoarticular infections, and they have shown successful results (30,31). Although no information is available regarding their effect on vertebral osteomyelitis specifically, it seems reasonable to assume that they work as well. Thus, oral fluorquinolones may be the best and most convenient therapy for PVO due to susceptible gram-negative bacilli, as was the case in 3 of our patients. The oral combination of rifampin with fluorquinolones, used in 1 of our cases, may be suitable for some patients with staphylococcal vertebral osteomyelitis. An 8-week treatment period is probably appropiate for most patients. The setting of vertebral osteomyelitis with both negative blood cultures and spinal needle biopsy challenges the clinician. In most cases, another needle biopsy or an open biopsy are recommended for microbiological diagnosis. Occasionally empir-
278
NOLLA ET AL
ical antibiotic therapy is prescribed, depending on the clinical characteristics of the patient. Relapse in PVO after appropriate medical therapy is infrequent. In our series, 3 of the 41 patients with long-term follow-up relapsed. Although the life outcome of PVO seems in general to be good (only 3% of our patients died
due to the sepsis), functional outcome is poorer. Sequelae, in the form of spinal pain or neurologic deficit, have been reported in 8% to 41% of subjects (16,17). The prognosis depends mainly on early diagnosis; therefore, efforts should be made to reduce the considerable diagnostic delay that occurs in patients with PVO.
REFERENCES 1. Lew DP, Waldvogel FA. Osteomyelitis. N Engl J Med 1997;336:999-1007. 2. Jime´nez-Mejı´as ME, Colmenero JD, Sa´nchez-Lora FJ, Palomino-Nica´s J, Reguera JM, Garcı´a de las Heras J et al. Postoperative spondylodiskitis: etiology, clinical findings, prognosis and comparison with nonoperative pyogenic spondylodiskitis. Clin Infect Dis 1999;29:339-45. 3. Ozuna RM, Delamarter RB. Pyogenic vertebral osteomyelitis and postsurgical disc space infections. Orthop Clin North Am 1996;27:87-93. 4. Sapico FL. Microbiology and antimicrobial therapy of spinal infections. Orthop Clin North Am 1996;27:9-13. 5. Sapico FL, Montgomerie JZ. Pyogenic vertebral osteomyelitis: report of nine cases and review of the literature. Rev Infect Dis 1979;1:754-76. 6. Osenbach RK, Hitchon PW, Menezes AH. Diagnosis and management of pyogenic vertebral osteomyelitis in adults. Surg Neurol 1990;33:266-75. 7. Patzakis MJ, Rao S, Wilkins J, Moore TM, Harvey PJ. Analysis of 61 cases of vertebral osteomyelitis. Clin Orthop 1991;264:178-83. 8. Meyers SP, Wiener SN. Diagnosis of hematogenous pyogenic vertebral osteomyelitis by magnetic resonance imaging. Arch Intern Med 1991;151:683-7. 9. Cahill DW, Love LC, Rechtine GR. Pyogenic osteomyelitis of the spine in the elderly. J Neurosurg 1991;74:878-86. 10. Bontoux D, Codello L, Debiais F, Lambert de Cursay G, Azais I, Alcalay M. Spondylodiscites infectieuses. Rev Rhum Mal Osteoartic 1992;59:401-7. 11. Buchelt M, Lack W, Kutschera HP, Katterschafka T, Kiss H, Schneider B, et al. Comparison of tuberculous and pyogenic spondylitis. Clin Orthop 1993;296:192-9. 12. Perronne C, Saba J, Behloul D, Salmon-Ce´ron D, Leport C, Vilde´ JL, et al. Pyogenic and tuberculous spondylodiskitis (vertebral osteomyelitis) in 80 adult patients. Clin Infect Dis 1994;19:746-50. 13. Torda AJ, Gottlieb T, Bradbury R. Pyogenic vertebral osteomyelitis: analysis of 20 cases and review. Clin Infect Dis 1995;20:320-8. 14. Honan M, White GW, Eisenberg GM. Spontaneous infectious discitis in adults. Am J Med 1996;100:85-9. 15. Fouquet B, Goupille P, Gobert F, Cotty P, Roulot B, Valat JP. Spondylodiscites infectieuses: apport diagnostique des examens biologiques et de la ponction-biopsie disco-verte´brale. Rev Rhum (Ed Fr) 1996;63:26-31. 16. Carragee EJ. Pyogenic vertebral osteomyelitis. J Bone Joint Surg Am 1997;79:874-80. 17. Colmenero JD, Jime´nez-Mejı´as ME, Sa´nchez-Lora FJ, Reguera JM, Palomino-Nica´s J, Martos F, et al. Pyogenic, tuberculous, and brucellar vertebral osteomyelitis: a descriptive
and comparative study of 219 cases. Ann Rheum Dis 1997;56: 709-15. 18. Pascaretti C, Legrand E, Degasne I, Masson C, Bre´geon C, Caron C, et al. Epidurites au cours des spondylodiscites infectieuses non tuberculeuses: fre´quence en IRM, signification, pronostic. Rev Rhum (Ed Fr) 1997;64:640-5. 19. Santos EM, Sapico FL. Vertebral osteomyelitis due to Salmonellae: report of two cases and review. Clin Infect Dis 1998;27:287-95. 20. Jensen AG, Espersen F, Skinhoj P, Fridmot-Moller N. Bacteremic Staphylococcus aureus spondylitis. Arch Intern Med 1998;158:509-17. 21. Turner DPJ, Weston VC, Ispahani P. Streptococcus pneumoniae spinal infection in Nottingham, United Kingdom: not a rare event. Clin Infect Dis 1999;28:873-81. 22. Veillard E, Guggenbuhl P, Morcet N, Meadeb J, Bello S, Perdriger A, et al. Re´gression rapide des abce`s paraverte´braux et des e´pidurites au cours de l’evolution des spondylodiscites a` germes banals. A propos de 16 spondylodiscites e´valu´ees en IRM. Rev Rhum (Ed Fr) 2000;67:219-28. 23. Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, eds. Manual of Clinical Microbiology (7th ed). Washington, DC. American Society for Microbiology 1999. 24. Belzunegui J, Del Val N, Intxausti JJ, De Dios JR, Queiro R, Gonza´lez C, et al. Vertebral osteomyelitis in northern Spain. Report of 62 cases. Clin Exp Rheumatol 1999;17:447-52. 25. Adatepe MH, Powell OM, Isaacs GH, Nichols K, Cefola R. Hematogenous pyogenic vertebral osteomyelitis: diagnostic value of radionuclide bone imaging. J Nucl Med 1986;27: 1680-5. 26. Abbey DM, Hosea SW. Diagnosis of vertebral osteomyelitis in a community hospital by using computed tomography. Arch Intern Med 1989;149:2029-35. 27. Oostveen JCM, van de Laar MAFJ. Magnetic resonance imaging in rheumatic disorders of the spine and sacroiliac joints. Semin Arthritis Rheum 2000;30:52-69. 28. Modic MT, Feiglin DH, Piraino DW, Boumphrey F, Weinstein MA, Duchesnau PM, et al. Vertebral osteomyelitis: assessment using MR. Radiology 1985;157:157-66. 29. Sullivan Baker A, Ojemann RG, Baker RA. Editorial response to Wheeler et al.: to decompress or not to decompress spinal epidural abscess. Clin Infect Dis 1992;15:28-9. 30. Lew DP, Waldvogel FA. Use of quinolones in osteomyelitis and infected orthopedic prosthesis. Drugs 1999;58(suppl 2):85-91. 31. Zimmerli W, Widmer AF, Blatter M, Frei R, Ochsner PE; for the foreign-body infection study group. Role of rifampicin for treatment of orthopedic implant-related staphylococcal infections. A randomized controlled trial. JAMA 1998;279:1537-41.
From the Rheumatology Department and Infectious Diseases Department, Ciutat Sanita`ria i Universita`ria de Bellvitge, Barcelona, Spain. Joan M. Nolla, MD, PhD: Professor of Medicine, Staff member, Rheumatology Department, Javier Ariza, MD, PhD: Professor of Medicine, Staff member, Infectious Diseases Department, Carmen Go´mez-Vaquero, MD, PhD: Staff member, Rheumatology Department, Jordi Fiter, MD: Staff member, Rheumatology Department, Joaquı´n Bermejo, MD, Fellow, Infectious Diseases Department, Josep Valverde, MD, PhD: Professor of Medicine, Staff member, Rheumatology Department, Daniel Roig Escofet, MD, PhD, Pro-
fessor of Medicine, Chair of Rheumatology Department, Francesc Gudiol, MD, PhD, Professor of Medicine, Chair of Infectious Diseases Department; Ciutat Sanita`ria i Universita`ria de Bellvitge, Barcelona, Spain. Address reprint requests to Dr. Joan M. Nolla, Rheumatology Service (Pl 10-2), Ciutat Sanita`ria i Universita`ria de Bellvitge, Feixa Llarga s/n, 08907 L⬘Hospitalet, Barcelona, Spain. E-mail address: [email protected] Copyright 2002, Elsevier Science (USA). All rights reserved. 0049-0172/02/3104-1048$35.00/0 doi:10.1053/sarh.2002.29492
Seminars in Arthritis and Rheumatism, Vol 31, No 4 (February), 2002: pp 271-278
271
272
NOLLA ET AL
P
YOGENIC VERTEBRAL osteomyelitis (PVO) has been regarded as an uncommon entity (1). Several investigators, however, have stressed that its incidence seems to be increasing as a result of the higher life expectancy of older patients with chronic debilitating diseases, the rise in the prevalence of intravenous drug abuse, and the increase in spinal instrumentations and surgery. A clear distinction should be made between postoperative and spontaneous PVO. Postoperative PVO is produced as the result of direct inoculation (2,3), whereas spontaneous PVO (3,4) is thought to be caused by hematogenous spread into the disc periphery or metaphysis of the vertebral bodies. The 2 forms have important clinical and microbiological differences and require different approaches. In spontaneous PVO (4), the clinical picture and relative prevalence of causative microorganisms are substantially different in drug and nondrug users. In drug users, the disease affects mainly young patients, follows a septic course, and Staphylococcus aureus is the etiologic agent in the great majority of cases; in these individuals, the presence of a strong risk factor makes the diagnosis easy. In nondrug users, PVO occurs predominantly in aged patients, septic manifestations are not always present, and a wide variety of gram-positive and gram-negative organisms act as causative agents; in this population, the diagnosis is easily missed. Until recently, the information available on PVO was limited (5). In the last 10 years, several series have been published (6-22); however, some were small (8,9,13,14) and others focused on 1 particular etiologic agent (19-21) or radiologic technique (8,18,22). As a result, certain diagnostic and therapeutic issues relating to the practical management of patients remain to be clarified. In this study, we review our experience with spontaneous PVO in nondrug users from the dual perspective of rheumatic and infectious disease specialists. We focus on decisive clinical and microbiological aspects of the disease and evaluate the contributions of modern radiology to the diagnostic and therapeutic management of patients. PATIENTS AND METHODS
Patients The charts of all adult patients in whom vertebral osteomyelitis was diagnosed between January
1980 and December 1999 in the Ciutat Sanita`ria i Universita`ria de Bellvitge, a 1,000-bed tertiary teaching hospital in the Barcelona area, were reviewed. Patients with prior spinal instrumentation or surgery, injection drug users, and those with tuberculosis or brucellosis were excluded. We included only patients in whom a definitive diagnosis of PVO was made by the presence of 3 criteria: 1) compatible clinical picture (spinal pain and localized tenderness or limited range of motion), 2) compatible imaging findings (decreased height of the intervertebral disk with or without erosions of the end plates on adjacent vertebral bodies on plain radiographs and/or increased uptake of tracers on technetium Tc 99m diphosphonate or gallium citrate Ga 67 bone scans), and 3) isolation of a pyogenic organism in blood cultures or spinal biopsy specimens. Sixty-four patients met the criteria; the cases were evenly distributed throughout the study period. A diagnosis of probable PVO was made in 11 patients (6 men and 5 women; mean age, 69 years) during the same period. They had a compatible clinical picture, suggestive imaging findings, and an appropriate response to antibiotic therapy, but no organism was isolated. Furthermore, 28 patients with tuberculous vertebral osteomyelitis and 35 patients with brucellar vertebral osteomyelitis were identified. Data Collection Every patient’s chart was retranscribed onto an standard itemized form and included the following information: 1. 2. 3. 4.
Demographic data Underlying medical illnesses Vertebral level involved Clinical presentation a. Duration of symptoms before diagnosis b. Type of spinal pain: mechanical (pain while moving) or inflammatory (pain at rest) c. Findings on spinal physical examination (tenderness, limited range of motion) d. Neurologic findings e. Maximal recorded temperature (°C) prior to diagnosis 5. Bacteriologic diagnosis 6. Presumed sources of infection and concomitant infectious processes
PYOGENIC VERTEBRAL OSTEOMYELITIS
7. Laboratory tests a. Erythrocyte sedimentation rate (ESR) evaluated by Westergren method b. White blood cell (WBC) count 8. Radiologic tests a. Plain radiography b. Technetium bone scan c. Gallium bone scan d. Computed tomography (CT) e. Magnetic resonance imaging (MRI) 9. Treatment and outcome Bacteriologic Studies A minimum of 2 blood samples were obtained from all patients before the start of antibiotic therapy. Spinal specimens were cultured in 5% horse blood agar, chocolate agar, and MacConkey agar plates and in thioglycolate medium. Microorganisms were identified according to standard criteria (23). Statistical Study The relation among categorical variables was established by the chi-square test. Correlations were made by means of Spearman’s test. Differences among groups of patients were assessed by analysis of variance. A P value of ⬍ .05 was considered significant. RESULTS
Demographic Data Of the 64 patients, 36 (56%) were men and 28 (44%) were women, with ages ranging from 18 to 84 years (mean, 59 ⫾ 17 years). Thirty-eight (59%) patients were over 60 years, and 7 (11%) were under 25. Underlying Medical Illnesses In 29 (45%) patients, 1 or more underlying medical illnesses were found: diabetes mellitus in 18, liver cirrhosis in 6, renal failure in 5, malignancies in 2, rheumatoid arthritis treated with glucocorticoids in 1, and congestive heart failure in 1. These patients were on average older than the 35 patients without relevant medical illnesses (65 ⫾ 8 years v 55 ⫾ 20 years; P ⬍ .05). Vertebral Level Involved Sixty-one (95%) patients had involvement of only 1 motion segment (2 adjacent vertebrae and
273
the intervening disc): the lumbar area in 40 cases, the thoracic area in 17, and the cervical area in 4. Three (5%) patients had involvement of several segments, which were contiguous in 2 cases (1 in lumbar area and 1 in thoracic area) and not contiguous in 1 (lumbar and cervical areas). Clinical Presentation The mean duration of symptoms before hospital admission was 48 days ⫾ 40 days. No significant correlation between the age of the patient and the duration of symptoms was found. No significant difference was found in the duration of symptoms among the etiologic agents. Back or neck pain was reported in all cases; it was inflammatory in 48 (75%) patients and mechanical in 16 (25%). Physical examination revealed localized spinal tenderness in 62 (97%) patients and a limited range of motion in 60 (94%); all patients had at least 1 of these findings. Neurologic impairment was present in 18 (28%) patients: 11 (17%) had radiculopathy, 5 (8%) had lower limb weakness, and 2 (3%) with cervical involvement due to S aureus had quadriparesis. Thirteen (20%) patients had an axillary temperature ⬍37°C. In 11 (17%) patients, the temperature ranged from 37°C to 38°C, and it was ⬎38°C in the remaining 40 (63%) patients. Bacteriologic Diagnosis In all cases, the infection was monomicrobial; the organisms isolated are listed in Table 1. S aureus and E coli were the predominant etiologic agents. S aureus was isolated in 26% (10/38) of patients over 60 years, in 50% (13/26) of those under 60 years, and in 100% (7/7) of those under 25 years. Seventy-three percent (11/15) of patients with PVO due to E coli were over 60 years of age. Table 2 shows the diagnostic procedures used. Etiologic diagnosis was established on the basis of blood cultures alone in 39 cases and on the basis of biopsy specimens alone in 18 cases (S aureus in 6 patients, Streptococcus agalactiae in 3, Pseudomonas aeruginosa in 2, and Streptococcus salivarius, Streptococcus mitis, Streptococcus sanguis, Haemophylus paraprophylus, Proteus vulgaris, Proteus mirabilis and Eikenella corrodens in 1 patient each). We identified the same organism in the 7 cases in which blood and spinal cultures gave positive results. Blood cultures were positive in 46 of 64 (72%)
274
NOLLA ET AL
Table 1: Organisms Isolated From 64 Patients With Pyogenic Vertebral Osteomyelitis
Bacteria
No. (%) of Cases
Staphylococci Staphylococcus aureus* Staphylococcus epidermidis Streptococci Streptococcus sanguis Streptococcus agalactiae Streptococcus salivarius Streptococcus mitis Streptococcus bovis Streptococcus pneumoniae Gram-negative bacilli Escherichia coli Pseudomonas aeruginosa Eikenella corrodens Proteus mirabilis Proteus vulgaris Haemophylus aprophylus Haemophylus paraprophylus Other organisms Enterococcus faecalis Bacteroides fragilis
25 (39) 23 2 12 (19) 5 3 1 1 1 1 25 (39) 15 3 2 2 1 1 1 2 (3) 1 1
*One case, methicillin-resistant (MRSA).
cases: 34 of the 40 (85%) patients with axillary temperature ⬎38°C, 6 of the 11 (54%) patients with axillary temperature 37°C to 38°C, and 6 of the 13 (46%) with axillary temperature ⬍37°C (2 of whom had endocarditis). Blood cultures were positive in 19 of 25 (76%) patients with staphylococcal infection or gram-negative bacterial infection and in 6 of 12 (50%) patients with streptococcal infection. Biopsy was performed in 36 patients. The specimen was obtained by percutaneous x-ray (usually CT) guided biopsy in 16 cases and by open biopsy in 15. In 5 patients, both biopsy techniques were performed. The yield of x-ray-guided biopsy was 52% (11/21) and of open biopsy was 75% (15/20). In 3 cases, in which needle biopsy and/or open biopsy were performed after antibiotic therapy had been initiated, the cultures of the obtained specimens were negative. Among the 5 cases with percutaneous and open biopsy, 1 gave positive results by both techniques; the cultures of the other 4 spinal specimens obtained by x-ray-guided bi-
opsy were negative, and 2 cultures obtained by open biopsy were positive. No complications related to biopsy were documented. Presumed Sources of Infection and Concomitant Infectious Processes PVO was community-acquired in all patients except for 1 hemodyalized older man, in whom the condition was secondary to a cannula-related sepsis due to methicillin-resistant S aureus. Overall, a source of infection or a concomitant infectious process was identified in 30 (47%) of patients. Lower urinary tract infection was the presumed source of infection in 13 patients, who developed compatible symptoms before the onset of back pain. E coli was the causative organism in 10 cases, P aeruginosa in 2, and P mirabilis in 1. In 2 patients, the urinary infection was proven microbiologically at the time of PVO diagnosis. Two patients with PVO due to S aureus had a cutaneous ulceration which was considered the source of infection. In 1 patient with PVO due to E coli, a colonic endoscopy with biopsy of a polyp had been performed 2 weeks before the onset of symptoms. One patient with Bacteroides fragilis infection had a periapendicular abscess. In the patient with Streptococcus bovis infection, colonic diverticulosis was diagnosed after the diagnosis of PVO. In 1 patient with infection due to E corrodens, a rectal neoplasm was diagnosed during the hospital stay. Two patients with cirrhosis had spontaneous peritonitis due to E coli. In 1 cirrhotic patient with PVO due to S epidermidis, an infection of a LeVeen shunt was identified. Blood cultures were positive in all cirrhotic patients.
Table 2: Diagnostic Procedures Used in 64 Cases With Pyogenic Vertebral Osteomyelitis
Diagnostic Procedures Blood cultures Blood cultures and x-ray–guided biopsy Blood cultures and open biopsy Blood cultures, x-ray–guided biopsy, and open biopsy
No. (%) of Patients 28 (44) 16 (25) 15 (23) 5 (8)
PYOGENIC VERTEBRAL OSTEOMYELITIS
Three patients with PVO due to S aureus presented other concomitant staphylococcal osteoarticular infections: knee arthritis, sacroiliitis, and elbow bursitis. Four patients suffered from endocarditis: S sanguis was the causative organism in 2 cases and E corrodens and S epidermidis in 1 each. Laboratory Tests At admission, the mean ESR (⫾ SE) was 77 mm/h ⫾ 33 mm/h (range, 8-140 mm/h); 6 (9%) patients had an ESR below 30 mm/h. The mean WBC count (⫾ SE) was 9.7 ⫾ 6 ⫻ 109/L (range, 1.7-37 ⫻ 109 /L); the WBC count was above 11 ⫻ 109/L in 23 (36%) patients. Radiologic Tests All patients underwent plain radiography of the spine, 53 (83%) had technetium bone scans, 44 (69%) had gallium scans, 36 (56%) had CT, and 33 (52%) had MRI. CT and MRI were performed only after plain radiography and bone scanning. CT and/or MRI were available in all cases with neurologic findings. Plain radiography revealed abnormalities in 57 of 64 patients. Increased uptake on technetium and/or gallium bone scans was observed in the 7 patients with negative x-rays; these patients had symptoms of short duration (mean, 14 days; range, 7-21 days). Fifty-one of 53 (96%) technetium and 41 of 44 (91%) gallium bone scans showed increased uptake of tracer in the affected vertebral level. In the 2 patients with negative technetium scanning and whose duration of disease was less than 15 days, gallium scans were positive. Among the 4 patients with negative gallium bone scans, all of whom presented pathologic findings on plain radiography, technetium scans were positive in 3 and scanning was not performed in 1. In 53 patients, a CT and/or an MRI was performed (CT alone in 20 cases, MRI alone in 17, and both in 16), and changes consistent with PVO were demonstrated in all cases. In 15 patients in whom a gadollinium-enhanced MRI was performed, contrast uptake was observed. Paraspinal or epidural extension was seen in 39 (74%) (paraspinal alone in 16 cases, epidural alone in 5 cases, and both 18). Compression of the spinal cord was identified in 7 cases, but no change in spinal cord signal was found. CT and MRI data
275
were the same. Diagnostic delay and etiologic agents were similar in patients with or without extra-vertebral extension. Fifteen of the 39 patients with extension had neurologic findings: 9 patients had radiculopathy (7 with paraspinal and 2 with both paraspinal and epidural extension), 5 patients had lower limb weakness (1 with paraspinal and 4 with both paraspinal and epidural extension), and 1 patient, with extensions into both areas, had quadriparesis. In 1 patient with quadriparesis, only signs of subluxation were observed on CT. Treatment and Outcome Four patients died. Two of them, with S aureus bacteremia, died of infection on days 4 and 8 after beginning appropiate antibiotic therapy. In the other 2 cases, death was not infection related but attributable to traumatic splenic rupture in 1 case on day 15 and myocardial infarction in the other on day 62. Fifty-seven patients received intravenous antibiotics during 4 to 6 weeks; 20 of them were given further oral therapy for 2 to 4 weeks. Four patients received only oral treatment for 8 weeks; 3 with PVO due to E coli were given fluoroquinolones and 1 with S aureus infection received a combination of rifampicin and ciprofloxacin. No antibioticrelated complications were noted. Bed rest was prescribed in all cases with external immobilization when movement was authorized. Surgery was indicated in 4 patients with severe neurologic findings. However, patients who had an open biopsy usually were drained during the surgery. The mean hospital stay was 57 ⫾ 37 days, and no differences were observed between patients with or without underlying diseases (65 ⫾ 47 days v 50 ⫾ 24 days; P ⫽ ns). Information regarding back pain at discharge was obtained in 50 cases: 16 patients had no pain, 30 had mechanical pain, and 4 had inflammatory pain. All patients with radiculopathy recovered. The 2 patients with quadriparesis had lower limb weakness, and 3 of the 4 surviving patients with lower limb weakness recovered; 1 had persistent symptoms. Forty-one patients were seen regularly in our outpatient clinics after discharge (mean follow-up, 68 ⫾ 50 weeks). Three patients who had been treated with intravenous antibiotics for 6 weeks relapsed. Two had diabetes with PVO due to
276
NOLLA ET AL
S aureus, and 1 patient with cirrhosis had P aeruginosa infection. In patients with staphylococcal infection, extravertebral extension had been identified by CT and/or MRI (paraspinal extension in 1 case and both paraspinal and epidural extension in the other); the patient with epidural extension had undergone surgical drainage. Relapses occurred at 6 and 5 months, respectively, after the end of antibiotic therapy. The patient with cirrhosis and P aeruginosa infection showed no extra-vertebral extension on CT. Relapse occurred 3 months after the end of antibiotic therapy. DISCUSSION
We present a large descriptive series of patients with vertebral osteomyelitis seen in our hospital for the last 20 years. Patients with prior spinal instrumentation or surgery and injection drug users were excluded, since both these population groups present specific differential characteristics. In addition, we excluded patients with brucellar or tuberculous vertebral osteomyelitis, which in our and in other Spanish investigators’ (24) experience accounts for 50% of cases. Furthermore, we analyzed only microbiologically proven cases, which represented 85% of our patients with supposed spontaneous PVO. All these aspects should be considered when comparing our results with those of other more heterogeneous series (6-12,15-17). The demographic characteristics of our sample were similar to those reported elsewhere, in which spontaneous PVO was found to affect mainly older men (6,8,10,13,14,16,17), who suffered from underlying medical illnesses in nearly half the cases (6,12,14,16,17). The medical conditions most frequently associated with PVO in our patients, as in those of other series (6,12,14,16,17,21) were liver cirrhosis, chronic renal failure, and primarily diabetes mellitus. The delay in diagnosing PVO in our series and in previous reports (6,12,13,17) is of particular concern. Patients with apyrexia or low-grade fever who have mechanical-type back pain are often misdiagnosed. But physical examination provides an important aid to diagnosis, since all of our patients had spinal tenderness or limited range of spinal motion, 2 findings closely related to PVO (3). Once a clinical diagnosis of vertebral osteomyelitis has been established, an attempt to identify
the causative organism should be made (1,3,4). Blood cultures provide microbiological diagnosis in many cases, since PVO is usually a monomicrobial disease and isolates from positive blood cultures can be assumed to be responsible for the process (4). No discrepancies between isolates from blood and spinal specimen cultures were found in the 7 patients in which both procedures provided positive results. The reported yield of blood cultures in PVO ranges from 42% to 82% (12,17), depending on the population included. In our series of only microbiologically proven cases, blood cultures provided a substantial yield, positive in 72% of all patients and in 85% of those with a temperature above 38°C. In some cases, bacteremia was detected before the onset of PVO and was associated with the primary source of infection. A notable finding was that 46% of our patients with apyrexia at admission had positive blood cultures at some time during the course of the disease. When blood cultures are negative, spinal specimens must be obtained by closed percutaneous or open surgical biopsy; needle biopsy is usually preferred because of safety considerations. In our study, cultures were positive in 52% and 75% of cases, respectively, figures similar to previous reports (6,10,12,14,15,17). The number of biopsies was relatively small, and only 5 patients had both biopsy techniques; thus, we cannot determine which type of biopsy was most sensitive. A large number of gram-positive and gramnegative organisms have been identified as causative agents in PVO (4). Overall, S aureus has been the most commonly isolated organism, accounting for more than 50% of reported cases. Gram-negative bacilli produce almost 30% of cases and streptococci around 10%, although these percentages vary depending on the population studied. Carragee (16) stressed the relevance of low-virulence organisms, usually considered contaminants, but other series have not corroborated this finding. Prospective studies are needed to determine whether these microorganisms are responsible for certain cases of PVO. Our results reveal certain clinical-bacteriological associations. S aureus accounted for a greater proportion of cases in young patients, whereas E coli was present more frequently in the aged. The proportion of streptococcal infections should be noted; these occurred mostly in patients with no
PYOGENIC VERTEBRAL OSTEOMYELITIS
recognized source of infection. The association of endocarditis and PVO, observed in 6% of our patients, has been reported in the literature infrequently. The possibility of concomitant endocarditis should be considered in patients with PVO, especially if Streptococcus spp is isolated from blood cultures. The diagnosis of PVO may require a number of imaging techniques (1,3). Conventional radiography is necessary at both presentation and follow-up (1). Plain radiographs show changes suggesting PVO in most patients at this time of clinical evaluation (6,8,13,17). Technetium diphosphonate and gallium citrate isotope scans (25) have high sensitivity but low specificity and may be useful in diagnosing early vertebral involvement in patients with normal plain radiography, as was the case in 7 of 64 (11%) of our cases. Additional information can be obtained when multiple noncontiguous vertebrae are involved, a rare circumstance in nondrug users (10,12-14). Technetium diphosphonate is probably the most suitable tracer because it is simpler and safer to use. CT is the technique of choice for the guidance of needle biopsies and provides useful information on bone changes, disc space narrowing, and soft tissue collections (26). MRI is an excellent technique for the diagnosis of vertebral osteomyelitis (1,27). Findings are seen early (28), and higher contrast resolution detects marrow infiltration and intradural disease (8,28) with sensitivity and specificity both above 90% (28). Unfortunately, MRI is expensive and not available in some centers. However, MRI should be performed if available. The information it provides, mainly on the presence of extra-vertebral extension, may be very useful if the clinical outcome during follow-up is unsatisfactory. In many cases, the diagnosis of PVO can be made on the basis of plain radiography, reserving radionuclide bone imaging techniques for initial evaluation of suspected cases with normal plain radiography and CT when needle biopsy is required. Paraspinal and/or epidural extension is a frequent finding when CT and MRI are used for the diagnosis of PVO (74% of the patients in our series). Like other authors (8), we used the term extension, since we did not attempt to differentiate diffuse inflammation from a true abscess, as that may be very difficult.
277
No relationship was found between extension, microorganism type, or other clinical findings. In fact, in most patients this radiologic finding had no clinical consequences. Twenty-eight percent of our patients with PVO had concomitant neurologic findings, a prevalence within the range previously reported (20% to 52%) (12,13). Radiculopathy was associated with paraspinal extension, and lower limb weakness or quadriparesis occurred with epidural extension. However, the presence of epidural extension was not necessarily associated with the development of neurologic findings. Medical treatment alone, including antibiotic therapy, bed rest, and immobilization with orthosis, is suitable for most cases of PVO; surgical intervention is reserved for the management of complications or for the failure of medical therapy (1,3). The main indications for surgery include open biopsy, spinal cord compression with significant neurologic deficit, large abscess, and substantial kyphotic deformity. The therapeutic approach should be based on the clinical manifestations and not on the imaging findings (29). The optimal duration of therapy is not defined, since nearly all reports of PVO are retrospective. Usually, antibiotics are given intravenously for at least 4 to 6 weeks, which is the main reason for the long hospital stay, and some clinicians use oral antibiotics for an additional several weeks (3). Recently, oral antibiotics as the only or main therapy has been increasingly recommended for different types of osteomyelitis and other serious osteoarticular infections, and they have shown successful results (30,31). Although no information is available regarding their effect on vertebral osteomyelitis specifically, it seems reasonable to assume that they work as well. Thus, oral fluorquinolones may be the best and most convenient therapy for PVO due to susceptible gram-negative bacilli, as was the case in 3 of our patients. The oral combination of rifampin with fluorquinolones, used in 1 of our cases, may be suitable for some patients with staphylococcal vertebral osteomyelitis. An 8-week treatment period is probably appropiate for most patients. The setting of vertebral osteomyelitis with both negative blood cultures and spinal needle biopsy challenges the clinician. In most cases, another needle biopsy or an open biopsy are recommended for microbiological diagnosis. Occasionally empir-
278
NOLLA ET AL
ical antibiotic therapy is prescribed, depending on the clinical characteristics of the patient. Relapse in PVO after appropriate medical therapy is infrequent. In our series, 3 of the 41 patients with long-term follow-up relapsed. Although the life outcome of PVO seems in general to be good (only 3% of our patients died
due to the sepsis), functional outcome is poorer. Sequelae, in the form of spinal pain or neurologic deficit, have been reported in 8% to 41% of subjects (16,17). The prognosis depends mainly on early diagnosis; therefore, efforts should be made to reduce the considerable diagnostic delay that occurs in patients with PVO.
REFERENCES 1. Lew DP, Waldvogel FA. Osteomyelitis. N Engl J Med 1997;336:999-1007. 2. Jime´nez-Mejı´as ME, Colmenero JD, Sa´nchez-Lora FJ, Palomino-Nica´s J, Reguera JM, Garcı´a de las Heras J et al. Postoperative spondylodiskitis: etiology, clinical findings, prognosis and comparison with nonoperative pyogenic spondylodiskitis. Clin Infect Dis 1999;29:339-45. 3. Ozuna RM, Delamarter RB. Pyogenic vertebral osteomyelitis and postsurgical disc space infections. Orthop Clin North Am 1996;27:87-93. 4. Sapico FL. Microbiology and antimicrobial therapy of spinal infections. Orthop Clin North Am 1996;27:9-13. 5. Sapico FL, Montgomerie JZ. Pyogenic vertebral osteomyelitis: report of nine cases and review of the literature. Rev Infect Dis 1979;1:754-76. 6. Osenbach RK, Hitchon PW, Menezes AH. Diagnosis and management of pyogenic vertebral osteomyelitis in adults. Surg Neurol 1990;33:266-75. 7. Patzakis MJ, Rao S, Wilkins J, Moore TM, Harvey PJ. Analysis of 61 cases of vertebral osteomyelitis. Clin Orthop 1991;264:178-83. 8. Meyers SP, Wiener SN. Diagnosis of hematogenous pyogenic vertebral osteomyelitis by magnetic resonance imaging. Arch Intern Med 1991;151:683-7. 9. Cahill DW, Love LC, Rechtine GR. Pyogenic osteomyelitis of the spine in the elderly. J Neurosurg 1991;74:878-86. 10. Bontoux D, Codello L, Debiais F, Lambert de Cursay G, Azais I, Alcalay M. Spondylodiscites infectieuses. Rev Rhum Mal Osteoartic 1992;59:401-7. 11. Buchelt M, Lack W, Kutschera HP, Katterschafka T, Kiss H, Schneider B, et al. Comparison of tuberculous and pyogenic spondylitis. Clin Orthop 1993;296:192-9. 12. Perronne C, Saba J, Behloul D, Salmon-Ce´ron D, Leport C, Vilde´ JL, et al. Pyogenic and tuberculous spondylodiskitis (vertebral osteomyelitis) in 80 adult patients. Clin Infect Dis 1994;19:746-50. 13. Torda AJ, Gottlieb T, Bradbury R. Pyogenic vertebral osteomyelitis: analysis of 20 cases and review. Clin Infect Dis 1995;20:320-8. 14. Honan M, White GW, Eisenberg GM. Spontaneous infectious discitis in adults. Am J Med 1996;100:85-9. 15. Fouquet B, Goupille P, Gobert F, Cotty P, Roulot B, Valat JP. Spondylodiscites infectieuses: apport diagnostique des examens biologiques et de la ponction-biopsie disco-verte´brale. Rev Rhum (Ed Fr) 1996;63:26-31. 16. Carragee EJ. Pyogenic vertebral osteomyelitis. J Bone Joint Surg Am 1997;79:874-80. 17. Colmenero JD, Jime´nez-Mejı´as ME, Sa´nchez-Lora FJ, Reguera JM, Palomino-Nica´s J, Martos F, et al. Pyogenic, tuberculous, and brucellar vertebral osteomyelitis: a descriptive
and comparative study of 219 cases. Ann Rheum Dis 1997;56: 709-15. 18. Pascaretti C, Legrand E, Degasne I, Masson C, Bre´geon C, Caron C, et al. Epidurites au cours des spondylodiscites infectieuses non tuberculeuses: fre´quence en IRM, signification, pronostic. Rev Rhum (Ed Fr) 1997;64:640-5. 19. Santos EM, Sapico FL. Vertebral osteomyelitis due to Salmonellae: report of two cases and review. Clin Infect Dis 1998;27:287-95. 20. Jensen AG, Espersen F, Skinhoj P, Fridmot-Moller N. Bacteremic Staphylococcus aureus spondylitis. Arch Intern Med 1998;158:509-17. 21. Turner DPJ, Weston VC, Ispahani P. Streptococcus pneumoniae spinal infection in Nottingham, United Kingdom: not a rare event. Clin Infect Dis 1999;28:873-81. 22. Veillard E, Guggenbuhl P, Morcet N, Meadeb J, Bello S, Perdriger A, et al. Re´gression rapide des abce`s paraverte´braux et des e´pidurites au cours de l’evolution des spondylodiscites a` germes banals. A propos de 16 spondylodiscites e´valu´ees en IRM. Rev Rhum (Ed Fr) 2000;67:219-28. 23. Murray PR, Baron EJ, Pfaller MA, Tenover FC, Yolken RH, eds. Manual of Clinical Microbiology (7th ed). Washington, DC. American Society for Microbiology 1999. 24. Belzunegui J, Del Val N, Intxausti JJ, De Dios JR, Queiro R, Gonza´lez C, et al. Vertebral osteomyelitis in northern Spain. Report of 62 cases. Clin Exp Rheumatol 1999;17:447-52. 25. Adatepe MH, Powell OM, Isaacs GH, Nichols K, Cefola R. Hematogenous pyogenic vertebral osteomyelitis: diagnostic value of radionuclide bone imaging. J Nucl Med 1986;27: 1680-5. 26. Abbey DM, Hosea SW. Diagnosis of vertebral osteomyelitis in a community hospital by using computed tomography. Arch Intern Med 1989;149:2029-35. 27. Oostveen JCM, van de Laar MAFJ. Magnetic resonance imaging in rheumatic disorders of the spine and sacroiliac joints. Semin Arthritis Rheum 2000;30:52-69. 28. Modic MT, Feiglin DH, Piraino DW, Boumphrey F, Weinstein MA, Duchesnau PM, et al. Vertebral osteomyelitis: assessment using MR. Radiology 1985;157:157-66. 29. Sullivan Baker A, Ojemann RG, Baker RA. Editorial response to Wheeler et al.: to decompress or not to decompress spinal epidural abscess. Clin Infect Dis 1992;15:28-9. 30. Lew DP, Waldvogel FA. Use of quinolones in osteomyelitis and infected orthopedic prosthesis. Drugs 1999;58(suppl 2):85-91. 31. Zimmerli W, Widmer AF, Blatter M, Frei R, Ochsner PE; for the foreign-body infection study group. Role of rifampicin for treatment of orthopedic implant-related staphylococcal infections. A randomized controlled trial. JAMA 1998;279:1537-41.