Infection DOI 10.1007/s15010-013-0582-0
CLINICAL AND EPIDEMIOLOGICAL STUDY
Outcome-related co-factors in 105 cases of vertebral osteomyelitis in a tertiary care hospital M. Loibl • L. Stoyanov • C. Doenitz • A. Brawanski • P. Wiggermann • W. Krutsch • M. Nerlich • M. Oszwald C. Neumann • B. Salzberger • F. Hanses
•
Received: 16 August 2013 / Accepted: 26 December 2013 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Vertebral osteomyelitis (VO) is an infection of the vertebral body and the adjacent disc space. The aim of our study was to identify outcome-related co-factors of patients with VO treated in the last decade. Methods and results We retrospectively identified 105 patients with VO (mean age 66.1 years) who had been treated at our institution from 2004 to 2011. The median time of hospitalization at our institution was 31.5 days, and 44 patients required intensive medical care. Back pain and fever were documented in 66.7 and 33.3 % of cases, respectively. The radiologic diagnosis of VO was made in 94.8 % of all obtained magnetic resonance imaging scans and in 66.2 % of all computed tomography (CT) scans. Biopsies were taken in 71 patients, and the causative organisms were identified in 56.2 % of patients, with Staphylococcus aureus being the predominant pathogen. Fifty-six patients underwent surgical treatment. During hospitalization, infectious complications were observed in M. Loibl W. Krutsch M. Nerlich M. Oszwald C. Neumann Department of Trauma Surgery, University Hospital Regensburg, Regensburg, Germany L. Stoyanov B. Salzberger F. Hanses (&) Department of Internal Medicine I, University Hospital Regensburg, Franz-Josef-Strauss-Allee 11, 93042 Regensburg, Germany e-mail: [email protected] C. Doenitz A. Brawanski Department of Neurosurgery, University Hospital Regensburg, Regensburg, Germany P. Wiggermann Department of Radiology, University Hospital Regensburg, Regensburg, Germany
63 patients (60.0 %). The most common complications were psoas, paravertebral and epidural abscesses. Patients with S. aureus infections had a significantly higher rate of infectious complications than those without (76.5 vs. 40.3 %, respectively), and were more frequently treated in intensive care units (58.8 vs. 34.7 %, respectively). Overall in-hospital mortality rate was 12.4 %. Elevated C-reactive protein levels at admission, advanced age and a Charlson Comorbidity Index of C2 were associated with higher mortality. Conclusions Magnetic resonance imaging currently is the imaging procedure of choice for the radiologic diagnosis of VO. Mortality is attributable in part to co-morbidities. However, infections with S. aureus are frequent in this patient population and are associated with a higher rate of complications and a trend towards higher mortality. Keywords Vertebral osteomyelitis Spondylodiscitis Outcome Spine Infection Staphylococcus aureus
Introduction Vertebral osteomyelitis (VO), or spondylodiscitis, is an infection of the vertebral body and the adjacent intervertebral disc space. Pathogens can reach the spine by hematogenous or contiguous spread from sites of infection or by direct inoculation from trauma or surgery. VO is the most frequent manifestation of hematogenous osteomyelitis in adult patients [1], and its annual incidence has been estimated to be between 1:42,000 and 1:450,000, with an increasing trend in recent years [2–4]. Gram-positive cocci, especially Staphylococcus aureus, are the most common microorganisms causing VO [5–11].
123
M. Loibl et al.
Several studies analyzed the clinical, microbiological, and radiologic features of VO [5, 8, 10–18]. Predisposing risk factors include advanced age [12], renal or liver disease [5], diabetes mellitus [16], malignant disease [5], immunosuppression [17, 19], intravenous drug abuse [9] and previous spinal surgery [13, 20]. Factors contributing to the rising incidence of VO may be an increasing life expectancy of patients, higher rates of chronic and immunosuppressive co-morbidities, intravenous drug abuse, an increasing use of instrumentation in spinal surgery, and increasing rates of nosocomial bacteremia [14]. The results of the most comprehensive retrospective analysis of 253 patients with a median follow-up of 6.5 years (range 2 days–38 years) need to be interpreted with caution since the cases were collected over a long time period, reaching back to the 1950s, and many cases occurred prior to the routine use of magnetic resonance imaging (MRI), which is the most sensitive radiologic technique for the diagnosis of VO [10]. Few reported series analyzed comprehensive epidemiological, diagnostic and clinical features of VO patients treated in the last decade. Therefore, certain clinical, diagnostic and therapeutic issues remain unclarified. To address this gap in our knowledge of VO, we reviewed records of patients with VO admitted to our hospital during an 8-year period.
Patients and methods We conducted a retrospective review of all clinical records of patients diagnosed with VO at the University Hospital of Regensburg (an 800-bed tertiary care medical center) between January 2004 and December 2011 using a standardized questionnaire. VO was defined by visualization of the radiologic features of VO by one or more imaging techniques (or positive microbiology and histology from biopsy specimen) in combination with clinical and/or laboratory signs of inflammation [fever, increased leucocytes or elevated C-reactive protein levels (CRP)] and the absence of other (inflammatory) spondylopathies. Items to be recorded included patient data [age, sex, body mass index, underlying co-morbidities, prior infectious diseases, immunosuppressive drugs, Charlson Comorbidity Index, Acute Physiology and Chronic Health Evaluation (APACHE) score, echocardiography, prior spine surgery, location of vertebral osteomyelitis (cervical, thoracic, lumbar), clinical course], diagnostic data {presence of fever and back pain, laboratory parameters [leucocytes, CRP], imaging [computed tomography (CT), MRI] results, biopsy (CT-guided, surgical) results and microbiology} and treatment data [treating discipline (e.g. neurosurgery, trauma surgery, internal medicine/infectious diseases), length of hospital stay, referral from another
123
institution, discharge to another institution, prior antibiotic treatment, antibiotic therapy, surgical treatment, 28-day mortality and hospital mortality]. Data were analyzed using SPSS ver. 16 (SPSS Inc., Chicago, IL). Kaplan–Meier estimates were used to calculate survival curves. The chi-square test and t test were used to calculate statistical significance where appropriate, and the Logrank test was used to compare survival distributions. Cox regression analyses were used to calculate statistical significance for multivariate analysis. P \ 0.05 was considered to be significant. The study was performed in concordance with the local Institutional Review Board’s requirements for retrospective research and did not require patients’ consent.
Results Patients A total of 114 patients with suspected VO were identified retrospectively. Of these, nine patients were excluded from the analysis: six patients had a history of previous VO only and were hospitalized for different reasons, one patient suffered from a paravertebral infection not involving the spine and one patient each had hematologic and rheumatologic disorders involving the spine without evidence of infection. The mean age of the remaining 105 patients was 66.1 ± 13.1 (range 28–88) years at the time of hospitalization. The study cohort comprised 55 men (52 %) and 50 women (48 %). Sixty-nine patients (65.7 %) had been hospitalized at another institution before referral to our hospital. Within our institution, 38 (36.2 %), 27 (25.7 %), and 40 (38.1 %) patients had been treated at the departments of neurosurgery, trauma surgery and internal medicine, respectively. The median time of hospitalization at our institution was 31.5 (range 2–198) days. The main reasons for a prolonged hospital stay were: (1) patients requiring more than one surgical intervention; (2) patients with multiple co-morbidities requiring medical attention; (3) no established outpatient intravenous antibiotic therapy; (4) a high rate of patients admitted to intensive care units (ICUs). Forty-four patients (41.9 %) required treatment at a surgical or medical ICU. Clinical and laboratory findings Back pain was the most common initial symptom and presented in 70 patients (66.7 %). Only 35 patients (33.3 %) presented with fever ([38.3 °C). Upon admission, an elevated leucocyte count was present in 72 patients (68.6 %); the mean leukocyte count was
Outcome-related co-factors in 105 cases
11.6 ± 4.3/nl. During hospitalization, the mean leucocyte count increased to a peak of 15.4 ± 6.3/nl and then decreased to 7.9 ± 3.9/nl at the time of discharge. The mean CRP values were 129.3 ± 105.1 mg/l upon admission, 187.2 ± 97.1 mg/l at maximum and 52.0 ± 60.3 mg/ l at the time of discharge. The radiologic diagnosis of VO was based on findings visualized on MRI, CT or positron emission tomography (PET) scans. MRI scans were obtained for 77 patients (73.3 %), CT scans for 77 (73.3 %) and PET scans for six (5.1 %). The radiologic diagnosis of VO was made in 73 of 77 MRI scans (94.8 %), in 51 of 77 CT scans (66.2 %), and in five of six PET scans (83.3 %). Among those patients with a negative MRI, one case of VO was identified via PET and three cases were confirmed intraoperatively later in the clinical course. The anatomic localization of VO is presented in Fig. 1. Echocardiography was performed in 61 patients (58.1 %), of whom 37 underwent transesophageal echocardiography. Endocarditis was found in seven patients (6.7 %). Causative organisms were identified in 59 patients (56.2 %). Percutaneous CT-guided or surgical biopsies were taken in 71 patients (67.6 %); of these, at least 17 biopsies were performed under antibiotic therapy (23.9 %). Cultures of the specimen obtained by percutaneous
Fig. 1 Localization of vertebral osteomyelitis (VO)
CT-guided biopsies were positive in 15 of 26 patients (57.7 %), and cultures of samples obtained during surgery were positive for 19 of 38 patients (50.0 %). For patients with negative cultures from percutaneous CT-guided biopsies, subsequent cultures of samples obtained during surgery were positive in four of seven cases (57.1 %). In 33 cases (46.5 %), cultures from biopsies were negative. Blood cultures were positive in 29 cases (i.e. 49 % of all cases with an identified pathogen), and among which cultures from biopsies were positive in 14 cases. Pathogens recovered from these 14 biopsies were identical in all but one case to those grown from the blood cultures. The distribution of the pathogens isolated is shown in Fig. 2. S. aureus was the predominant pathogen found in 34 patients (32.4 %), of whom five had methicillin-resistant S. aureus (MRSA). Coagulase-negative staphylococci (CoNS) were isolated in 12 patients (11.4 %)—in nine cases from the biopsy specimen and in three cases from repeated blood cultures. Among the patients with VO due to CoNS, three had previous spinal surgery, one patient had a proven intravascular catheter infection, three patients had other foreign bodies implanted (knee and hip prostheses), two were on immunosuppressive therapy after solid organ transplantation and one patient had both S. aureus and S. epidermidis in the biopsy. The remaining pathogens were streptococcal (7 patients, 6.7 %), enterococcal (4 patients, 3.8 %) and Candida species (4 patients, 3.8 %). In eight patients other pathogens could be isolated, including Mycobacterium tuberculosis (39), Acinetobacter, Enterobacteriaceae, Pseudomonas aeruginosa and Pasteurella multocida. Patients with S. aureus infections presented with higher leukocyte counts than those without S. aureus infections or with infections caused by other pathogens (13.36 ± 0.78 vs. 10.72 ± 0.48/nl, respectively; P = 0.003), as well as with higher CRP levels at admission (169.7 ± 18.4 vs.
Fig. 2 Causative microorganisms in 105 cases of VO. CoNS Coagulase-negative staphylococci
123
M. Loibl et al. Table 1 Patient demographics, underlying co-morbidities and infectious complications in patients with vertebral osteomyelitis Patients Male Age, mean (years) Body mass index, mean (kg/m2)
n = 105 (100 %) 55 (52 %) 66.1 ± 13.1 (28–88) 24.7 ± 6.5 (15–45)
Co-morbidities Coronary artery disease or heart failure
41 (39.0 %)
Renal failure (acute and chronic)
36 (34.3 %)
Diabetes mellitus Previous spinal surgery
35 (33.3 %) 20 (19.0 %)
Rheumatoid arthritis or osteoporosis
12 (11.4 %)
Malignancy
11 (10.5 %)
Neurological disorders
11 (10.5 %)
Respiratory disorders
10 (9.5 %)
Liver cirrhosis
10 (9.5 %)
Vascular disease
9 (8.6 %)
Gastrointestinal disorders
6 (5.7 %)
Skin disorders
4 (3.8 %)
Endocrine disorders
1 (1.0 %)
Human immunodeficiency virus
1 (1.0 %)
Infectious complications Psoas and paravertebral abscess
31 (29.5 %)
Epidural abscess
19 (18.1 %)
Pleura empyema Joint empyema
7 (6.7 %) 7 (6.7 %)
Septic encephalopathy
4 (3.8 %)
Infected intravascular devices
3 (2.9 %)
Subcutaneous abscesses
3 (2.9 %)
Endocarditis
2 (1.9 %)
Liver and spleen abscesses
1 (1.0 %)
Data are presented as the mean ± standard deviation and, where appropriate, with the range in parenthesis
108.5 ± 12.1 mg/l, respectively; P = 0.005). We also observed a trend towards a predominance of male patients in the S. aureus-infected subgroups only [20 of 34 patients (58.8 %) with VO due to S. aureus were male compared to 35 of the 71 remaining patients (49.3 %); difference is not significant]. Co-morbidities and outcome One or more comorbidities were present in 87 of 105 patients (82.6 %). Coronary artery disease or heart failure were the most common diseases and were present in 41 patients (39.0 %). Renal failure (acute and chronic) and diabetes mellitus were present in 36 (34.3 %) and 35 (33.3 %) patients, respectively (Table 1). Fourteen patients were receiving immunosuppressive drug therapy (13.3 %). Fortyfive patients had no previously documented infection
123
(42.9 %), and 70 patients had at least one documented site of infection, among whom nine patients had more than two documented sites of infection in the body. The most commonly found co-infections were infections of the upper respiratory tract (22.9 %), urogenital tract (14.3 %), skin and subcutaneous tissues (14.3 %), bone/joints (7.6 %) and gastrointestinal tract (6.7 %), sepsis (4.8 %), endocarditis (4.8 %) and meningitis (3.8 %). Surgical treatment was performed for 56 of 105 patients (53.3 %). Drainage of abscesses followed by spinal decompression was performed in eight patients (7.6 %), spinal stabilization in six patients (5.7 %), vertebral body fusion in 16 patients (15.2 %) and corpectomy and replacement of the vertebral body in 26 patients (24.8 %). In 12 patients more than one surgery was necessary (11.4 %); 20 patients (19.0 %) had undergone previous spinal surgery. Patients treated on ICUs were more likely to require a surgical intervention than those who were not (67.4 vs. 42.4 %, respectively; P = 0.018), but they also had significantly higher inflammation parameters on admission [leukocyte count 13.01 vs. 10.48/nl, respectively (P = 0.003); CRP 156.8 vs. 106.8 mg/l, respectively (P = 0.017)]. At least 23 patients (21.9 %) who had previous antibiotic treatment were admitted to our hospital. The overall pathogen recovery rate was similar among patients in this group. We observed a trend towards fewer pathogens cultivated from blood cultures in these 23 patients (21.7 vs. 28 %). However, none of these differences were significant. Within our institution, all patients received intravenous antibiotic treatment. The most commonly used antibiotics were flucloxacillin, ciprofloxacin, vancomycin and rifampicin, followed by piperacillin/tazobactam, clindamycin and ceftriaxone. The most commonly used empiric therapy was a combination of betalactams with either rifampicin or ciprofloxacin. During hospitalization, infectious complications [including psoas abscesses (25 patients), paravertebral abscesses (6 patients, 29.5 % of total) and epidural abscesses (19 patients, 18.1 % of total)] were detected or occurred in 63 patients (60.0 %) (Table 1). Neurologic complications developed in 29 patients (27.6 %), motor weakness or paralysis was observed in 23 patients (21.9 %), persisting sensory deficits were found in three patients (2.9 %), liquor fistulas were observed in two patients (1.9 %) and general reduced consciousness could be observed in one patient (1.0 %) with septic encephalopathy. Patients with S. aureus infections had a significantly higher rate of infectious complications than those with no or other pathogens (mainly abscess formation: 76.5 vs. 40.3 %, respectively; P = 0.002) and were more frequently treated in ICUs (58.8 vs. 34.7 %, respectively; P = 0.019). We also observed a non-significant trend towards higher mortality in patients with S. aureus infections (Fig. 3a).
Outcome-related co-factors in 105 cases
Fig. 3 Cumulative survival (Kaplan–Meier survival curves) based on infection with Staphylococcus aureus (a), C-reactive protein at admission (b), age [60 years (c) and a Charlson Comorbidity Index of C2 (d). 60y 60 years of age, n.s. not significant
Thirty-four patients (32.4 %) were discharged home after recovery, 29 patients were discharged to in-patient rehabilitation centers (27.6 %) and 29 patients were discharged to the referring hospital (27.6 %) for continuation of therapy. The in-hospital mortality rate was 12.4 % (13 patients). CRP levels of C100 mg/l at admission, age of C60 years and a Charlson Comorbidity Index [21] of C2 were associated with a higher mortality (Fig. 3). No significant differences in mortality were found in multivariate analysis (Cox regression).
Discussion We retrospectively examined patients with VO treated at our institution during an 8-year period (January 2004–
December 2011). Of the 105 included patients, the diagnosis of pyogenic VO was confirmed microbiologically in 59 patients (56.2 %). Most previous studies included about 100 patients with VO [4, 7, 9, 12, 22, 23], whereas one multi-center study [10] involving seven hospitals included 253 patients. Some of these studies are quite old, and a substantial portion of the patients involved were recruited up to 60 years ago. In contrast, our study gives a comprehensive analysis of epidemiological, diagnostic and clinical features of patients with VO treated in the last decade. Several studies reported predisposing risk factors for VO, such as diabetes mellitus, which is a prominent risk factor [16], and advanced age [12], renal or liver disease [5], intravenous drug abuse [9], malignant disease [5], immunosuppression [19] and previous spinal surgery [13,
123
M. Loibl et al.
20]. The majority of our patients had a variety of comorbidities, with the most frequent being cardiac disease, renal failure, diabetes mellitus and osteoporosis/rheumatoid arthritis. McHenry et al. [10] found that the severity of co-morbidities was not an independent risk factor for an adverse outcome. However, a Charlson Comorbidity Index (quantifying the severity of co-morbidities based on 22 different conditions and thus correlating with 10-year mortality, [21]) of C2 was associated with higher mortality in our study. Similarly, an age of C60 years was associated with a higher mortality. Mortality in patients with VO thus seems to be driven mainly by the overall state of health of the patient. Previous studies found that men are affected from VO approximately twice as often as women, with a male to female ratio of 1.5–2:1 [2, 5, 24]. In our cohort, gender was more evenly distributed, and a trend towards a male predominance was observed in the subgroup with S. aureus infections only. The reasons for this distorted gender distribution in VO and in invasive S. aureus infections in general remain to be determined. An early diagnosis of VO is difficult to establish. The most common presenting symptoms reported in previous reviews [10, 25] were fever and back pain. In our study, fever and back pain was identified in 33.3 and 66.7 % of the patients, respectively. The number of patients with back pain is especially low ([90 % in previous studies) and might be attributed to the retrospective nature of the study. In addition, the use of antipyretic drugs in patients who are referred to a tertiary care hospital might also mask the clinical presentation of fever. McHenry et al. [10] demonstrated that diagnostic delay is an independent risk factor for unfavorable outcome. We found that CRP levels of C100 mg/l at admission were associated with a higher mortality in patients with VO. Khan et al. [26] similarly demonstrated the value of CRP for following treatment response to antibiotics in wound infections after spinal surgery. Only one patient had a CRP value within the normal range upon admission; however, this patient was admitted for coronary artery bypass surgery initially and developed VO later in the clinical course. In a previous study, an increased leukocyte count was associated with poor sensitivity for the diagnosis and prognosis of VO [22]; in our study it was not associated with higher mortality of VO patients. MRI is considered to be the modality of choice for the radiologic diagnosis of VO with a sensitivity of 96 % [27]. Its advantage over other modalities lies within its superior ability to provide information about soft tissues, including the neural tissues and abscesses. We found a comparable sensitivity of 94.8 %. Despite the fact that the timely use of MRI shortens the time to diagnosis, MRI scans were obtained for 77 patients (73.3 %) only. Unfortunately, we
123
cannot report on the time to diagnosis of VO in our patients. Despite the decreased sensitivity of CT scans in the diagnosis of VO, CT scans were obtained in 77 patients (73.3 %), possibly due to the fact that percutaneous CTguided biopsies were taken in 26 patients (24.8 %) in the same session or to CT scans being performed in preparation for spinal surgery. The value of percutaneous CTguided biopsies as a safe and minimally invasive intervention is well established [28]. Although several biopsies were taken when the patient was on antibiotic treatment, cultures of 57.7 % of the specimens obtained by percutaneous CT-guided biopsies were positive, and cultures of 50 % of the samples obtained during surgery were positive. Gras et al [29]. demonstrated a slightly lower diagnostic yield of percutaneous needle biopsies (43.4 %) in patients with VO. Some clinicians would consider a negative percutaneous biopsy result as an indication for surgical biopsy [30]. In our study, four out of seven negative cultures from percutaneous CT-guided biopsies were positive in subsequent cultures of samples obtained during surgery (57.1 %). Although the total number of open biopsies after negative percutaneous biopsy was low in our study, their yield seemed to be higher than the 39.4 % positivity rate reported for a second percutaneous biopsy [29]. A higher diagnostic yield had been previously attributed to open biopsies rather than percutaneous needle biopsies. However, this higher diagnostic yield might also have been due to differences in the patient population, with more neurological complications, higher inflammation parameters and potentially higher bacterial burden in patients requiring surgery [10, 31]. The overall pathogen recovery rate is low in our study. However, our study cohort most likely represents a preselected patient population. The reasons patients were chosen for referral might include medical, surgical and diagnostic uncertainties, and many patients may have received (disclosed or undisclosed) antimicrobial therapies prior to referral. A wide range of bacterial, mycobacterial, fungal and parasitic organisms have been associated with VO. However, S. aureus currently represents the predominant pathogen in 36–62 % of patients with VO in developed countries [9, 10, 12, 23]. We isolated S. aureus in 32.4 % of our patients (49.3 % of those with an identified pathogen). Five of these strains were MRSA (14.7 %), which is in the range of the local MRSA prevalence. Park et al. [32] performed a retrospective cohort study in Asia in patients with S. aureus VO and found a higher percentage of MRSA—44.6 % of the cases. Moreover, these authors reported that MRSA was associated with a higher risk of persistent bacteremia, relapse and increased hospital stay. In our study we could not confirm a worse outcome for patients with MRSA infections; however, this parameter might be underestimated due to the low number of MRSA
Outcome-related co-factors in 105 cases
infections among our patient cohort. CoNS were isolated from 12 patients with VO. Patients with infections due to CoNS had extensive previous healthcare exposure, and all but one had been referred from outside hospitals. Most of these patients had either previous spinal surgery or a foreign body implanted or were on immunosuppressive therapy—VO due to CoNS without documented relevant risk factors was a rare event and we cannot exclude previous possible sources of infection in the remaining patients. Most recent studies [10, 31, 33] estimate the percentage of CoNS at 6.6–10 %, with a trend towards higher numbers in more recent studies. The percentage of patients with CoNS in our study might be higher due to the fact that our cohort from a referral center is most likely highly preselected. Finally, in contrast to other studies [25], only few Gramnegative bacteria were isolated from patients with VO in our cohort. Since the advent of antibiotics, mortality due to VO has dropped from 56 % [34]. The 12.4 % in-hospital mortality for patients with VO reported here is comparable to the 0–16 % reported in other studies during the last decades [7, 12, 14]. Infection-related mortality of VO patients has been reported to be associated with S. aureus bacteremia [11]. We observed a trend towards higher mortality in patients with VO caused by S. aureus, but the total number of cases may have been too small to determine significant differences. However, we observed that these patients suffered from a higher rate of infectious complications (mainly abscess formation) and had to be treated in ICUs more frequently than patients with no or other pathogens, thereby highlighting the virulence of S. aureus in the setting of vertebral infections. Increased morbidity in VO caused by S. aureus (together with a wide range of potential pathogens and the rising antimicrobial resistance [35]) underlines the importance of ascertaining a definitive microbiological diagnosis. Due to its retrospective nature, our study has a number of inherent limitations. Patients were identified using hospital records and were double-checked against bacteriological records. There may have been a patient-selection bias, and treatment algorithms varied among physicians, departments and referring hospitals. There were no standard operating procedures for the treatment of VO, including surgery. ICU admission rates were high in our study, and many patients admitted to the ICUs underwent surgery; our institution does not have intermediate care units for these patients. However, patients admitted to ICUs also had significantly higher inflammation parameters upon admission, i.e. before surgery, and higher rates of infections due to S. aureus. The administration of antibiotics varied with regard to both type and duration. However, some of these limitations represent the current clinical reality. In summary, in-hospital mortality of
patients with VO is mainly attributable to age and comorbidities. However, infections with S. aureus are frequent and are associated with a higher rate of complications and a trend towards higher mortality. Early diagnosis of VO and microbiological detection of the responsible pathogen remain a challenge, especially in cohorts with preselected patients, such as in referral centers. Acknowledgments We would like to thank the Publishing Support from AO Foundation, Davos, Switzerland for the image of Fig. 1, which was adapted with permission. Conflict of interest
None.
References 1. Gouliouris T, Aliyu SH, Brown NM. Spondylodiscitis: update on diagnosis and management. J Antimicrob Chemother. 2010;65[Suppl 3]:iii11–24. doi:10.1093/jac/dkq303. 2. Grammatico L, Baron S, Rusch E, Lepage B, Surer N, Desenclos JC, et al. Epidemiology of vertebral osteomyelitis (VO) in France: analysis of hospital-discharge data 2002–2003. Epidemiol Infect. 2008;136:653–60. doi:10.1017/S0950268807008850. 3. Digby JM, Kersley JB. Pyogenic non-tuberculous spinal infection: an analysis of thirty cases. J Bone Joint Surg Br. 1979; 61:47–55. 4. Beronius M, Bergman B, Andersson R. Vertebral osteomyelitis in Goteborg, Sweden: a retrospective study of patients during 1990–1995. Scand J Infect Dis. 2001;33:527–32. 5. Mylona E, Samarkos M, Kakalou E, Fanourgiakis P, Skoutelis A. Pyogenic vertebral osteomyelitis: a systematic review of clinical characteristics. Semin Arthritis Rheum. 2009;39:10–7. doi:10. 1016/j.semarthrit.2008.03.002. 6. Lew DP, Waldvogel FA. Osteomyelitis. Lancet. 2004;364: 369–79. doi:10.1016/S0140-6736(04)16727-5. 7. Nolla JM, Ariza J, Gomez-Vaquero C, Fiter J, Bermejo J, Valverde J, et al. Spontaneous pyogenic vertebral osteomyelitis in nondrug users. Semin Arthritis Rheum. 2002;31:271–8. 8. Zimmerli W. Clinical practice. Vertebral osteomyelitis. N Engl J Med. 2010;362:1022–9. doi:10.1056/NEJMcp0910753. 9. Patzakis MJ, Rao S, Wilkins J, Moore TM, Harvey PJ. Analysis of 61 cases of vertebral osteomyelitis. Clin Orthop Relat Res. 1991;264:178–83. 10. McHenry MC, Easley KA, Locker GA. Vertebral osteomyelitis: long-term outcome for 253 patients from 7 Cleveland-area hospitals. Clin Infect Dis. 2002;34:1342–50. doi:10.1086/340102. 11. Pigrau C, Almirante B, Flores X, Falco V, Rodriguez D, Gasser I, et al. Spontaneous pyogenic vertebral osteomyelitis and endocarditis: incidence, risk factors, and outcome. Am J Med. 2005;118:1287. doi:10.1016/j.amjmed.2005.02.027. 12. Carragee EJ. Pyogenic vertebral osteomyelitis. J Bone Joint Surg Am. 1997;79:874–80. 13. Dufour V, Feydy A, Rillardon L, Redondo A, Le Page L, Bert F, et al. Comparative study of postoperative and spontaneous pyogenic spondylodiscitis. Semin Arthritis Rheum. 2005;34:766–71. doi:10.1016/j.semarthrit.2004.08.004. 14. Fantoni M, Trecarichi EM, Rossi B, Mazzotta V, Di Giacomo G, Nasto LA, et al. Epidemiological and clinical features of pyogenic spondylodiscitis. Eur Rev Med Pharmacol Sci. 2012;16:2–7. 15. Pola E, Logroscino CA, Gentiempo M, Colangelo D, Mazzotta V, Di Meco E, et al. Medical and surgical treatment of pyogenic spondylodiscitis. Eur Rev Med Pharmacol Sci. 2012;16:35–49.
123
M. Loibl et al. 16. Krogsgaard MR, Wagn P, Bengtsson J. Epidemiology of acute vertebral osteomyelitis in Denmark: 137 cases in Denmark 1978–1982, compared to cases reported to the National Patient Register 1991–1993. Acta Orthop Scand. 1998;69:513–7. 17. Rezai AR, Woo HH, Errico TJ, Cooper PR. Contemporary management of spinal osteomyelitis. Neurosurgery. 1999;44: 1018–25 (discussion 25–6). 18. Furitsch M, Trager K, van der Linden M, Spellerberg B. Group A streptococcal vertebral osteomyelitis presenting with acute quadriplegia. Infection. 2011;39:389–91. doi:10.1007/s15010011-0113-9. 19. Weinstein MA, Eismont FJ. Infections of the spine in patients with human immunodeficiency virus. J Bone Joint Surg Am. 2005;87:604–9. doi:10.2106/JBJS.C.01062. 20. Yuste JR, Alfonso M, Bustos C, Quintana J, Rubio M, Villas C, et al. Iliac bone Candida albicans osteomyelitis in a patient with iliac crest bone autograft: a case report and review of the literature. Infection. 2012;40:445–9. doi:10.1007/s15010-012-0276-z. 21. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40: 373–83. 22. Jensen AG, Espersen F, Skinhoj P, Frimodt-Moller N. Bacteremic Staphylococcus aureus spondylitis. Arch Intern Med. 1998;158:509–17. 23. Hadjipavlou AG, Mader JT, Necessary JT, Muffoletto AJ. Hematogenous pyogenic spinal infections and their surgical management. Spine (Phila Pa 1976). 2000;25:1668–79. 24. Sapico FL, Montgomerie JZ. Pyogenic vertebral osteomyelitis: report of nine cases and review of the literature. Rev Infect Dis. 1979;1:754–76. 25. Graham SM, Fishlock A, Millner P, Sandoe J. The management of Gram-negative bacterial haematogenous vertebral osteomyelitis: a case series of diagnosis, treatment and therapeutic outcomes. Eur Spine J. 2013;22:1845–53. doi:10.1007/s00586-0132750-4. 26. Khan MH, Smith PN, Rao N, Donaldson WF. Serum C-reactive protein levels correlate with clinical response in patients treated
123
27.
28.
29.
30. 31.
32.
33.
34. 35.
with antibiotics for wound infections after spinal surgery. Spine J. 2006;6:311–5. doi:10.1016/j.spinee.2005.07.006. Modic MT, Feiglin DH, Piraino DW, Boumphrey F, Weinstein MA, Duchesneau PM, et al. Vertebral osteomyelitis: assessment using MR. Radiology. 1985;157:157–66. Enoch DA, Cargill JS, Laing R, Herbert S, Corrah TW, Brown NM. Value of CT-guided biopsy in the diagnosis of septic discitis. J Clin Pathol. 2008;61:750–3. doi:10.1136/jcp.2007. 054296. Gras G, Buzele R, Parienti JJ, Debiais F, Dinh A, Dupon M, et al. Microbiological diagnosis of vertebral osteomyelitis: relevance of second percutaneous biopsy following initial negative biopsy and limited yield of post-biopsy blood cultures. Eur J Clin Microbiol Infect Dis. 2013;. doi:10.1007/s10096-013-1965-y. Sapico FL, Montgomerie JZ. Vertebral osteomyelitis. Infect Dis Clin North Am. 1990;4:539–50. Marschall J, Bhavan KP, Olsen MA, Fraser VJ, Wright NM, Warren DK. The impact of prebiopsy antibiotics on pathogen recovery in hematogenous vertebral osteomyelitis. Clin Infect Dis. 2011;52:867–72. doi:10.1093/cid/cir062. Park KH, Chong YP, Kim SH, Lee SO, Choi SH, Lee MS, et al. Clinical characteristics and therapeutic outcomes of hematogenous vertebral osteomyelitis caused by methicillin-resistant Staphylococcus aureus. J Infect. 2013;67:556–64. doi:10.1016/j. jinf.2013.07.026. Bhavan KP, Marschall J, Olsen MA, Fraser VJ, Wright NM, Warren DK. The epidemiology of hematogenous vertebral osteomyelitis: a cohort study in a tertiary care hospital. BMC Infect Dis. 2010;10:158. doi:10.1186/1471-2334-10-158. Bauman GI, Stifel RE. Osteomyelitis of the spine. Ann Surg. 1923;78:119–21. Lillie P, Thaker H, Moss P, Baruah J, Cullen L, Taylor D, et al. Healthcare associated discitis in the era of antimicrobial resistance. J Clin Rheumatol. 2008;14:234–7. doi:10.1097/RHU. 0b013e318181addd.
CLINICAL AND EPIDEMIOLOGICAL STUDY
Outcome-related co-factors in 105 cases of vertebral osteomyelitis in a tertiary care hospital M. Loibl • L. Stoyanov • C. Doenitz • A. Brawanski • P. Wiggermann • W. Krutsch • M. Nerlich • M. Oszwald C. Neumann • B. Salzberger • F. Hanses
•
Received: 16 August 2013 / Accepted: 26 December 2013 Ó Springer-Verlag Berlin Heidelberg 2014
Abstract Purpose Vertebral osteomyelitis (VO) is an infection of the vertebral body and the adjacent disc space. The aim of our study was to identify outcome-related co-factors of patients with VO treated in the last decade. Methods and results We retrospectively identified 105 patients with VO (mean age 66.1 years) who had been treated at our institution from 2004 to 2011. The median time of hospitalization at our institution was 31.5 days, and 44 patients required intensive medical care. Back pain and fever were documented in 66.7 and 33.3 % of cases, respectively. The radiologic diagnosis of VO was made in 94.8 % of all obtained magnetic resonance imaging scans and in 66.2 % of all computed tomography (CT) scans. Biopsies were taken in 71 patients, and the causative organisms were identified in 56.2 % of patients, with Staphylococcus aureus being the predominant pathogen. Fifty-six patients underwent surgical treatment. During hospitalization, infectious complications were observed in M. Loibl W. Krutsch M. Nerlich M. Oszwald C. Neumann Department of Trauma Surgery, University Hospital Regensburg, Regensburg, Germany L. Stoyanov B. Salzberger F. Hanses (&) Department of Internal Medicine I, University Hospital Regensburg, Franz-Josef-Strauss-Allee 11, 93042 Regensburg, Germany e-mail: [email protected] C. Doenitz A. Brawanski Department of Neurosurgery, University Hospital Regensburg, Regensburg, Germany P. Wiggermann Department of Radiology, University Hospital Regensburg, Regensburg, Germany
63 patients (60.0 %). The most common complications were psoas, paravertebral and epidural abscesses. Patients with S. aureus infections had a significantly higher rate of infectious complications than those without (76.5 vs. 40.3 %, respectively), and were more frequently treated in intensive care units (58.8 vs. 34.7 %, respectively). Overall in-hospital mortality rate was 12.4 %. Elevated C-reactive protein levels at admission, advanced age and a Charlson Comorbidity Index of C2 were associated with higher mortality. Conclusions Magnetic resonance imaging currently is the imaging procedure of choice for the radiologic diagnosis of VO. Mortality is attributable in part to co-morbidities. However, infections with S. aureus are frequent in this patient population and are associated with a higher rate of complications and a trend towards higher mortality. Keywords Vertebral osteomyelitis Spondylodiscitis Outcome Spine Infection Staphylococcus aureus
Introduction Vertebral osteomyelitis (VO), or spondylodiscitis, is an infection of the vertebral body and the adjacent intervertebral disc space. Pathogens can reach the spine by hematogenous or contiguous spread from sites of infection or by direct inoculation from trauma or surgery. VO is the most frequent manifestation of hematogenous osteomyelitis in adult patients [1], and its annual incidence has been estimated to be between 1:42,000 and 1:450,000, with an increasing trend in recent years [2–4]. Gram-positive cocci, especially Staphylococcus aureus, are the most common microorganisms causing VO [5–11].
123
M. Loibl et al.
Several studies analyzed the clinical, microbiological, and radiologic features of VO [5, 8, 10–18]. Predisposing risk factors include advanced age [12], renal or liver disease [5], diabetes mellitus [16], malignant disease [5], immunosuppression [17, 19], intravenous drug abuse [9] and previous spinal surgery [13, 20]. Factors contributing to the rising incidence of VO may be an increasing life expectancy of patients, higher rates of chronic and immunosuppressive co-morbidities, intravenous drug abuse, an increasing use of instrumentation in spinal surgery, and increasing rates of nosocomial bacteremia [14]. The results of the most comprehensive retrospective analysis of 253 patients with a median follow-up of 6.5 years (range 2 days–38 years) need to be interpreted with caution since the cases were collected over a long time period, reaching back to the 1950s, and many cases occurred prior to the routine use of magnetic resonance imaging (MRI), which is the most sensitive radiologic technique for the diagnosis of VO [10]. Few reported series analyzed comprehensive epidemiological, diagnostic and clinical features of VO patients treated in the last decade. Therefore, certain clinical, diagnostic and therapeutic issues remain unclarified. To address this gap in our knowledge of VO, we reviewed records of patients with VO admitted to our hospital during an 8-year period.
Patients and methods We conducted a retrospective review of all clinical records of patients diagnosed with VO at the University Hospital of Regensburg (an 800-bed tertiary care medical center) between January 2004 and December 2011 using a standardized questionnaire. VO was defined by visualization of the radiologic features of VO by one or more imaging techniques (or positive microbiology and histology from biopsy specimen) in combination with clinical and/or laboratory signs of inflammation [fever, increased leucocytes or elevated C-reactive protein levels (CRP)] and the absence of other (inflammatory) spondylopathies. Items to be recorded included patient data [age, sex, body mass index, underlying co-morbidities, prior infectious diseases, immunosuppressive drugs, Charlson Comorbidity Index, Acute Physiology and Chronic Health Evaluation (APACHE) score, echocardiography, prior spine surgery, location of vertebral osteomyelitis (cervical, thoracic, lumbar), clinical course], diagnostic data {presence of fever and back pain, laboratory parameters [leucocytes, CRP], imaging [computed tomography (CT), MRI] results, biopsy (CT-guided, surgical) results and microbiology} and treatment data [treating discipline (e.g. neurosurgery, trauma surgery, internal medicine/infectious diseases), length of hospital stay, referral from another
123
institution, discharge to another institution, prior antibiotic treatment, antibiotic therapy, surgical treatment, 28-day mortality and hospital mortality]. Data were analyzed using SPSS ver. 16 (SPSS Inc., Chicago, IL). Kaplan–Meier estimates were used to calculate survival curves. The chi-square test and t test were used to calculate statistical significance where appropriate, and the Logrank test was used to compare survival distributions. Cox regression analyses were used to calculate statistical significance for multivariate analysis. P \ 0.05 was considered to be significant. The study was performed in concordance with the local Institutional Review Board’s requirements for retrospective research and did not require patients’ consent.
Results Patients A total of 114 patients with suspected VO were identified retrospectively. Of these, nine patients were excluded from the analysis: six patients had a history of previous VO only and were hospitalized for different reasons, one patient suffered from a paravertebral infection not involving the spine and one patient each had hematologic and rheumatologic disorders involving the spine without evidence of infection. The mean age of the remaining 105 patients was 66.1 ± 13.1 (range 28–88) years at the time of hospitalization. The study cohort comprised 55 men (52 %) and 50 women (48 %). Sixty-nine patients (65.7 %) had been hospitalized at another institution before referral to our hospital. Within our institution, 38 (36.2 %), 27 (25.7 %), and 40 (38.1 %) patients had been treated at the departments of neurosurgery, trauma surgery and internal medicine, respectively. The median time of hospitalization at our institution was 31.5 (range 2–198) days. The main reasons for a prolonged hospital stay were: (1) patients requiring more than one surgical intervention; (2) patients with multiple co-morbidities requiring medical attention; (3) no established outpatient intravenous antibiotic therapy; (4) a high rate of patients admitted to intensive care units (ICUs). Forty-four patients (41.9 %) required treatment at a surgical or medical ICU. Clinical and laboratory findings Back pain was the most common initial symptom and presented in 70 patients (66.7 %). Only 35 patients (33.3 %) presented with fever ([38.3 °C). Upon admission, an elevated leucocyte count was present in 72 patients (68.6 %); the mean leukocyte count was
Outcome-related co-factors in 105 cases
11.6 ± 4.3/nl. During hospitalization, the mean leucocyte count increased to a peak of 15.4 ± 6.3/nl and then decreased to 7.9 ± 3.9/nl at the time of discharge. The mean CRP values were 129.3 ± 105.1 mg/l upon admission, 187.2 ± 97.1 mg/l at maximum and 52.0 ± 60.3 mg/ l at the time of discharge. The radiologic diagnosis of VO was based on findings visualized on MRI, CT or positron emission tomography (PET) scans. MRI scans were obtained for 77 patients (73.3 %), CT scans for 77 (73.3 %) and PET scans for six (5.1 %). The radiologic diagnosis of VO was made in 73 of 77 MRI scans (94.8 %), in 51 of 77 CT scans (66.2 %), and in five of six PET scans (83.3 %). Among those patients with a negative MRI, one case of VO was identified via PET and three cases were confirmed intraoperatively later in the clinical course. The anatomic localization of VO is presented in Fig. 1. Echocardiography was performed in 61 patients (58.1 %), of whom 37 underwent transesophageal echocardiography. Endocarditis was found in seven patients (6.7 %). Causative organisms were identified in 59 patients (56.2 %). Percutaneous CT-guided or surgical biopsies were taken in 71 patients (67.6 %); of these, at least 17 biopsies were performed under antibiotic therapy (23.9 %). Cultures of the specimen obtained by percutaneous
Fig. 1 Localization of vertebral osteomyelitis (VO)
CT-guided biopsies were positive in 15 of 26 patients (57.7 %), and cultures of samples obtained during surgery were positive for 19 of 38 patients (50.0 %). For patients with negative cultures from percutaneous CT-guided biopsies, subsequent cultures of samples obtained during surgery were positive in four of seven cases (57.1 %). In 33 cases (46.5 %), cultures from biopsies were negative. Blood cultures were positive in 29 cases (i.e. 49 % of all cases with an identified pathogen), and among which cultures from biopsies were positive in 14 cases. Pathogens recovered from these 14 biopsies were identical in all but one case to those grown from the blood cultures. The distribution of the pathogens isolated is shown in Fig. 2. S. aureus was the predominant pathogen found in 34 patients (32.4 %), of whom five had methicillin-resistant S. aureus (MRSA). Coagulase-negative staphylococci (CoNS) were isolated in 12 patients (11.4 %)—in nine cases from the biopsy specimen and in three cases from repeated blood cultures. Among the patients with VO due to CoNS, three had previous spinal surgery, one patient had a proven intravascular catheter infection, three patients had other foreign bodies implanted (knee and hip prostheses), two were on immunosuppressive therapy after solid organ transplantation and one patient had both S. aureus and S. epidermidis in the biopsy. The remaining pathogens were streptococcal (7 patients, 6.7 %), enterococcal (4 patients, 3.8 %) and Candida species (4 patients, 3.8 %). In eight patients other pathogens could be isolated, including Mycobacterium tuberculosis (39), Acinetobacter, Enterobacteriaceae, Pseudomonas aeruginosa and Pasteurella multocida. Patients with S. aureus infections presented with higher leukocyte counts than those without S. aureus infections or with infections caused by other pathogens (13.36 ± 0.78 vs. 10.72 ± 0.48/nl, respectively; P = 0.003), as well as with higher CRP levels at admission (169.7 ± 18.4 vs.
Fig. 2 Causative microorganisms in 105 cases of VO. CoNS Coagulase-negative staphylococci
123
M. Loibl et al. Table 1 Patient demographics, underlying co-morbidities and infectious complications in patients with vertebral osteomyelitis Patients Male Age, mean (years) Body mass index, mean (kg/m2)
n = 105 (100 %) 55 (52 %) 66.1 ± 13.1 (28–88) 24.7 ± 6.5 (15–45)
Co-morbidities Coronary artery disease or heart failure
41 (39.0 %)
Renal failure (acute and chronic)
36 (34.3 %)
Diabetes mellitus Previous spinal surgery
35 (33.3 %) 20 (19.0 %)
Rheumatoid arthritis or osteoporosis
12 (11.4 %)
Malignancy
11 (10.5 %)
Neurological disorders
11 (10.5 %)
Respiratory disorders
10 (9.5 %)
Liver cirrhosis
10 (9.5 %)
Vascular disease
9 (8.6 %)
Gastrointestinal disorders
6 (5.7 %)
Skin disorders
4 (3.8 %)
Endocrine disorders
1 (1.0 %)
Human immunodeficiency virus
1 (1.0 %)
Infectious complications Psoas and paravertebral abscess
31 (29.5 %)
Epidural abscess
19 (18.1 %)
Pleura empyema Joint empyema
7 (6.7 %) 7 (6.7 %)
Septic encephalopathy
4 (3.8 %)
Infected intravascular devices
3 (2.9 %)
Subcutaneous abscesses
3 (2.9 %)
Endocarditis
2 (1.9 %)
Liver and spleen abscesses
1 (1.0 %)
Data are presented as the mean ± standard deviation and, where appropriate, with the range in parenthesis
108.5 ± 12.1 mg/l, respectively; P = 0.005). We also observed a trend towards a predominance of male patients in the S. aureus-infected subgroups only [20 of 34 patients (58.8 %) with VO due to S. aureus were male compared to 35 of the 71 remaining patients (49.3 %); difference is not significant]. Co-morbidities and outcome One or more comorbidities were present in 87 of 105 patients (82.6 %). Coronary artery disease or heart failure were the most common diseases and were present in 41 patients (39.0 %). Renal failure (acute and chronic) and diabetes mellitus were present in 36 (34.3 %) and 35 (33.3 %) patients, respectively (Table 1). Fourteen patients were receiving immunosuppressive drug therapy (13.3 %). Fortyfive patients had no previously documented infection
123
(42.9 %), and 70 patients had at least one documented site of infection, among whom nine patients had more than two documented sites of infection in the body. The most commonly found co-infections were infections of the upper respiratory tract (22.9 %), urogenital tract (14.3 %), skin and subcutaneous tissues (14.3 %), bone/joints (7.6 %) and gastrointestinal tract (6.7 %), sepsis (4.8 %), endocarditis (4.8 %) and meningitis (3.8 %). Surgical treatment was performed for 56 of 105 patients (53.3 %). Drainage of abscesses followed by spinal decompression was performed in eight patients (7.6 %), spinal stabilization in six patients (5.7 %), vertebral body fusion in 16 patients (15.2 %) and corpectomy and replacement of the vertebral body in 26 patients (24.8 %). In 12 patients more than one surgery was necessary (11.4 %); 20 patients (19.0 %) had undergone previous spinal surgery. Patients treated on ICUs were more likely to require a surgical intervention than those who were not (67.4 vs. 42.4 %, respectively; P = 0.018), but they also had significantly higher inflammation parameters on admission [leukocyte count 13.01 vs. 10.48/nl, respectively (P = 0.003); CRP 156.8 vs. 106.8 mg/l, respectively (P = 0.017)]. At least 23 patients (21.9 %) who had previous antibiotic treatment were admitted to our hospital. The overall pathogen recovery rate was similar among patients in this group. We observed a trend towards fewer pathogens cultivated from blood cultures in these 23 patients (21.7 vs. 28 %). However, none of these differences were significant. Within our institution, all patients received intravenous antibiotic treatment. The most commonly used antibiotics were flucloxacillin, ciprofloxacin, vancomycin and rifampicin, followed by piperacillin/tazobactam, clindamycin and ceftriaxone. The most commonly used empiric therapy was a combination of betalactams with either rifampicin or ciprofloxacin. During hospitalization, infectious complications [including psoas abscesses (25 patients), paravertebral abscesses (6 patients, 29.5 % of total) and epidural abscesses (19 patients, 18.1 % of total)] were detected or occurred in 63 patients (60.0 %) (Table 1). Neurologic complications developed in 29 patients (27.6 %), motor weakness or paralysis was observed in 23 patients (21.9 %), persisting sensory deficits were found in three patients (2.9 %), liquor fistulas were observed in two patients (1.9 %) and general reduced consciousness could be observed in one patient (1.0 %) with septic encephalopathy. Patients with S. aureus infections had a significantly higher rate of infectious complications than those with no or other pathogens (mainly abscess formation: 76.5 vs. 40.3 %, respectively; P = 0.002) and were more frequently treated in ICUs (58.8 vs. 34.7 %, respectively; P = 0.019). We also observed a non-significant trend towards higher mortality in patients with S. aureus infections (Fig. 3a).
Outcome-related co-factors in 105 cases
Fig. 3 Cumulative survival (Kaplan–Meier survival curves) based on infection with Staphylococcus aureus (a), C-reactive protein at admission (b), age [60 years (c) and a Charlson Comorbidity Index of C2 (d). 60y 60 years of age, n.s. not significant
Thirty-four patients (32.4 %) were discharged home after recovery, 29 patients were discharged to in-patient rehabilitation centers (27.6 %) and 29 patients were discharged to the referring hospital (27.6 %) for continuation of therapy. The in-hospital mortality rate was 12.4 % (13 patients). CRP levels of C100 mg/l at admission, age of C60 years and a Charlson Comorbidity Index [21] of C2 were associated with a higher mortality (Fig. 3). No significant differences in mortality were found in multivariate analysis (Cox regression).
Discussion We retrospectively examined patients with VO treated at our institution during an 8-year period (January 2004–
December 2011). Of the 105 included patients, the diagnosis of pyogenic VO was confirmed microbiologically in 59 patients (56.2 %). Most previous studies included about 100 patients with VO [4, 7, 9, 12, 22, 23], whereas one multi-center study [10] involving seven hospitals included 253 patients. Some of these studies are quite old, and a substantial portion of the patients involved were recruited up to 60 years ago. In contrast, our study gives a comprehensive analysis of epidemiological, diagnostic and clinical features of patients with VO treated in the last decade. Several studies reported predisposing risk factors for VO, such as diabetes mellitus, which is a prominent risk factor [16], and advanced age [12], renal or liver disease [5], intravenous drug abuse [9], malignant disease [5], immunosuppression [19] and previous spinal surgery [13,
123
M. Loibl et al.
20]. The majority of our patients had a variety of comorbidities, with the most frequent being cardiac disease, renal failure, diabetes mellitus and osteoporosis/rheumatoid arthritis. McHenry et al. [10] found that the severity of co-morbidities was not an independent risk factor for an adverse outcome. However, a Charlson Comorbidity Index (quantifying the severity of co-morbidities based on 22 different conditions and thus correlating with 10-year mortality, [21]) of C2 was associated with higher mortality in our study. Similarly, an age of C60 years was associated with a higher mortality. Mortality in patients with VO thus seems to be driven mainly by the overall state of health of the patient. Previous studies found that men are affected from VO approximately twice as often as women, with a male to female ratio of 1.5–2:1 [2, 5, 24]. In our cohort, gender was more evenly distributed, and a trend towards a male predominance was observed in the subgroup with S. aureus infections only. The reasons for this distorted gender distribution in VO and in invasive S. aureus infections in general remain to be determined. An early diagnosis of VO is difficult to establish. The most common presenting symptoms reported in previous reviews [10, 25] were fever and back pain. In our study, fever and back pain was identified in 33.3 and 66.7 % of the patients, respectively. The number of patients with back pain is especially low ([90 % in previous studies) and might be attributed to the retrospective nature of the study. In addition, the use of antipyretic drugs in patients who are referred to a tertiary care hospital might also mask the clinical presentation of fever. McHenry et al. [10] demonstrated that diagnostic delay is an independent risk factor for unfavorable outcome. We found that CRP levels of C100 mg/l at admission were associated with a higher mortality in patients with VO. Khan et al. [26] similarly demonstrated the value of CRP for following treatment response to antibiotics in wound infections after spinal surgery. Only one patient had a CRP value within the normal range upon admission; however, this patient was admitted for coronary artery bypass surgery initially and developed VO later in the clinical course. In a previous study, an increased leukocyte count was associated with poor sensitivity for the diagnosis and prognosis of VO [22]; in our study it was not associated with higher mortality of VO patients. MRI is considered to be the modality of choice for the radiologic diagnosis of VO with a sensitivity of 96 % [27]. Its advantage over other modalities lies within its superior ability to provide information about soft tissues, including the neural tissues and abscesses. We found a comparable sensitivity of 94.8 %. Despite the fact that the timely use of MRI shortens the time to diagnosis, MRI scans were obtained for 77 patients (73.3 %) only. Unfortunately, we
123
cannot report on the time to diagnosis of VO in our patients. Despite the decreased sensitivity of CT scans in the diagnosis of VO, CT scans were obtained in 77 patients (73.3 %), possibly due to the fact that percutaneous CTguided biopsies were taken in 26 patients (24.8 %) in the same session or to CT scans being performed in preparation for spinal surgery. The value of percutaneous CTguided biopsies as a safe and minimally invasive intervention is well established [28]. Although several biopsies were taken when the patient was on antibiotic treatment, cultures of 57.7 % of the specimens obtained by percutaneous CT-guided biopsies were positive, and cultures of 50 % of the samples obtained during surgery were positive. Gras et al [29]. demonstrated a slightly lower diagnostic yield of percutaneous needle biopsies (43.4 %) in patients with VO. Some clinicians would consider a negative percutaneous biopsy result as an indication for surgical biopsy [30]. In our study, four out of seven negative cultures from percutaneous CT-guided biopsies were positive in subsequent cultures of samples obtained during surgery (57.1 %). Although the total number of open biopsies after negative percutaneous biopsy was low in our study, their yield seemed to be higher than the 39.4 % positivity rate reported for a second percutaneous biopsy [29]. A higher diagnostic yield had been previously attributed to open biopsies rather than percutaneous needle biopsies. However, this higher diagnostic yield might also have been due to differences in the patient population, with more neurological complications, higher inflammation parameters and potentially higher bacterial burden in patients requiring surgery [10, 31]. The overall pathogen recovery rate is low in our study. However, our study cohort most likely represents a preselected patient population. The reasons patients were chosen for referral might include medical, surgical and diagnostic uncertainties, and many patients may have received (disclosed or undisclosed) antimicrobial therapies prior to referral. A wide range of bacterial, mycobacterial, fungal and parasitic organisms have been associated with VO. However, S. aureus currently represents the predominant pathogen in 36–62 % of patients with VO in developed countries [9, 10, 12, 23]. We isolated S. aureus in 32.4 % of our patients (49.3 % of those with an identified pathogen). Five of these strains were MRSA (14.7 %), which is in the range of the local MRSA prevalence. Park et al. [32] performed a retrospective cohort study in Asia in patients with S. aureus VO and found a higher percentage of MRSA—44.6 % of the cases. Moreover, these authors reported that MRSA was associated with a higher risk of persistent bacteremia, relapse and increased hospital stay. In our study we could not confirm a worse outcome for patients with MRSA infections; however, this parameter might be underestimated due to the low number of MRSA
Outcome-related co-factors in 105 cases
infections among our patient cohort. CoNS were isolated from 12 patients with VO. Patients with infections due to CoNS had extensive previous healthcare exposure, and all but one had been referred from outside hospitals. Most of these patients had either previous spinal surgery or a foreign body implanted or were on immunosuppressive therapy—VO due to CoNS without documented relevant risk factors was a rare event and we cannot exclude previous possible sources of infection in the remaining patients. Most recent studies [10, 31, 33] estimate the percentage of CoNS at 6.6–10 %, with a trend towards higher numbers in more recent studies. The percentage of patients with CoNS in our study might be higher due to the fact that our cohort from a referral center is most likely highly preselected. Finally, in contrast to other studies [25], only few Gramnegative bacteria were isolated from patients with VO in our cohort. Since the advent of antibiotics, mortality due to VO has dropped from 56 % [34]. The 12.4 % in-hospital mortality for patients with VO reported here is comparable to the 0–16 % reported in other studies during the last decades [7, 12, 14]. Infection-related mortality of VO patients has been reported to be associated with S. aureus bacteremia [11]. We observed a trend towards higher mortality in patients with VO caused by S. aureus, but the total number of cases may have been too small to determine significant differences. However, we observed that these patients suffered from a higher rate of infectious complications (mainly abscess formation) and had to be treated in ICUs more frequently than patients with no or other pathogens, thereby highlighting the virulence of S. aureus in the setting of vertebral infections. Increased morbidity in VO caused by S. aureus (together with a wide range of potential pathogens and the rising antimicrobial resistance [35]) underlines the importance of ascertaining a definitive microbiological diagnosis. Due to its retrospective nature, our study has a number of inherent limitations. Patients were identified using hospital records and were double-checked against bacteriological records. There may have been a patient-selection bias, and treatment algorithms varied among physicians, departments and referring hospitals. There were no standard operating procedures for the treatment of VO, including surgery. ICU admission rates were high in our study, and many patients admitted to the ICUs underwent surgery; our institution does not have intermediate care units for these patients. However, patients admitted to ICUs also had significantly higher inflammation parameters upon admission, i.e. before surgery, and higher rates of infections due to S. aureus. The administration of antibiotics varied with regard to both type and duration. However, some of these limitations represent the current clinical reality. In summary, in-hospital mortality of
patients with VO is mainly attributable to age and comorbidities. However, infections with S. aureus are frequent and are associated with a higher rate of complications and a trend towards higher mortality. Early diagnosis of VO and microbiological detection of the responsible pathogen remain a challenge, especially in cohorts with preselected patients, such as in referral centers. Acknowledgments We would like to thank the Publishing Support from AO Foundation, Davos, Switzerland for the image of Fig. 1, which was adapted with permission. Conflict of interest
None.
References 1. Gouliouris T, Aliyu SH, Brown NM. Spondylodiscitis: update on diagnosis and management. J Antimicrob Chemother. 2010;65[Suppl 3]:iii11–24. doi:10.1093/jac/dkq303. 2. Grammatico L, Baron S, Rusch E, Lepage B, Surer N, Desenclos JC, et al. Epidemiology of vertebral osteomyelitis (VO) in France: analysis of hospital-discharge data 2002–2003. Epidemiol Infect. 2008;136:653–60. doi:10.1017/S0950268807008850. 3. Digby JM, Kersley JB. Pyogenic non-tuberculous spinal infection: an analysis of thirty cases. J Bone Joint Surg Br. 1979; 61:47–55. 4. Beronius M, Bergman B, Andersson R. Vertebral osteomyelitis in Goteborg, Sweden: a retrospective study of patients during 1990–1995. Scand J Infect Dis. 2001;33:527–32. 5. Mylona E, Samarkos M, Kakalou E, Fanourgiakis P, Skoutelis A. Pyogenic vertebral osteomyelitis: a systematic review of clinical characteristics. Semin Arthritis Rheum. 2009;39:10–7. doi:10. 1016/j.semarthrit.2008.03.002. 6. Lew DP, Waldvogel FA. Osteomyelitis. Lancet. 2004;364: 369–79. doi:10.1016/S0140-6736(04)16727-5. 7. Nolla JM, Ariza J, Gomez-Vaquero C, Fiter J, Bermejo J, Valverde J, et al. Spontaneous pyogenic vertebral osteomyelitis in nondrug users. Semin Arthritis Rheum. 2002;31:271–8. 8. Zimmerli W. Clinical practice. Vertebral osteomyelitis. N Engl J Med. 2010;362:1022–9. doi:10.1056/NEJMcp0910753. 9. Patzakis MJ, Rao S, Wilkins J, Moore TM, Harvey PJ. Analysis of 61 cases of vertebral osteomyelitis. Clin Orthop Relat Res. 1991;264:178–83. 10. McHenry MC, Easley KA, Locker GA. Vertebral osteomyelitis: long-term outcome for 253 patients from 7 Cleveland-area hospitals. Clin Infect Dis. 2002;34:1342–50. doi:10.1086/340102. 11. Pigrau C, Almirante B, Flores X, Falco V, Rodriguez D, Gasser I, et al. Spontaneous pyogenic vertebral osteomyelitis and endocarditis: incidence, risk factors, and outcome. Am J Med. 2005;118:1287. doi:10.1016/j.amjmed.2005.02.027. 12. Carragee EJ. Pyogenic vertebral osteomyelitis. J Bone Joint Surg Am. 1997;79:874–80. 13. Dufour V, Feydy A, Rillardon L, Redondo A, Le Page L, Bert F, et al. Comparative study of postoperative and spontaneous pyogenic spondylodiscitis. Semin Arthritis Rheum. 2005;34:766–71. doi:10.1016/j.semarthrit.2004.08.004. 14. Fantoni M, Trecarichi EM, Rossi B, Mazzotta V, Di Giacomo G, Nasto LA, et al. Epidemiological and clinical features of pyogenic spondylodiscitis. Eur Rev Med Pharmacol Sci. 2012;16:2–7. 15. Pola E, Logroscino CA, Gentiempo M, Colangelo D, Mazzotta V, Di Meco E, et al. Medical and surgical treatment of pyogenic spondylodiscitis. Eur Rev Med Pharmacol Sci. 2012;16:35–49.
123
M. Loibl et al. 16. Krogsgaard MR, Wagn P, Bengtsson J. Epidemiology of acute vertebral osteomyelitis in Denmark: 137 cases in Denmark 1978–1982, compared to cases reported to the National Patient Register 1991–1993. Acta Orthop Scand. 1998;69:513–7. 17. Rezai AR, Woo HH, Errico TJ, Cooper PR. Contemporary management of spinal osteomyelitis. Neurosurgery. 1999;44: 1018–25 (discussion 25–6). 18. Furitsch M, Trager K, van der Linden M, Spellerberg B. Group A streptococcal vertebral osteomyelitis presenting with acute quadriplegia. Infection. 2011;39:389–91. doi:10.1007/s15010011-0113-9. 19. Weinstein MA, Eismont FJ. Infections of the spine in patients with human immunodeficiency virus. J Bone Joint Surg Am. 2005;87:604–9. doi:10.2106/JBJS.C.01062. 20. Yuste JR, Alfonso M, Bustos C, Quintana J, Rubio M, Villas C, et al. Iliac bone Candida albicans osteomyelitis in a patient with iliac crest bone autograft: a case report and review of the literature. Infection. 2012;40:445–9. doi:10.1007/s15010-012-0276-z. 21. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40: 373–83. 22. Jensen AG, Espersen F, Skinhoj P, Frimodt-Moller N. Bacteremic Staphylococcus aureus spondylitis. Arch Intern Med. 1998;158:509–17. 23. Hadjipavlou AG, Mader JT, Necessary JT, Muffoletto AJ. Hematogenous pyogenic spinal infections and their surgical management. Spine (Phila Pa 1976). 2000;25:1668–79. 24. Sapico FL, Montgomerie JZ. Pyogenic vertebral osteomyelitis: report of nine cases and review of the literature. Rev Infect Dis. 1979;1:754–76. 25. Graham SM, Fishlock A, Millner P, Sandoe J. The management of Gram-negative bacterial haematogenous vertebral osteomyelitis: a case series of diagnosis, treatment and therapeutic outcomes. Eur Spine J. 2013;22:1845–53. doi:10.1007/s00586-0132750-4. 26. Khan MH, Smith PN, Rao N, Donaldson WF. Serum C-reactive protein levels correlate with clinical response in patients treated
123
27.
28.
29.
30. 31.
32.
33.
34. 35.
with antibiotics for wound infections after spinal surgery. Spine J. 2006;6:311–5. doi:10.1016/j.spinee.2005.07.006. Modic MT, Feiglin DH, Piraino DW, Boumphrey F, Weinstein MA, Duchesneau PM, et al. Vertebral osteomyelitis: assessment using MR. Radiology. 1985;157:157–66. Enoch DA, Cargill JS, Laing R, Herbert S, Corrah TW, Brown NM. Value of CT-guided biopsy in the diagnosis of septic discitis. J Clin Pathol. 2008;61:750–3. doi:10.1136/jcp.2007. 054296. Gras G, Buzele R, Parienti JJ, Debiais F, Dinh A, Dupon M, et al. Microbiological diagnosis of vertebral osteomyelitis: relevance of second percutaneous biopsy following initial negative biopsy and limited yield of post-biopsy blood cultures. Eur J Clin Microbiol Infect Dis. 2013;. doi:10.1007/s10096-013-1965-y. Sapico FL, Montgomerie JZ. Vertebral osteomyelitis. Infect Dis Clin North Am. 1990;4:539–50. Marschall J, Bhavan KP, Olsen MA, Fraser VJ, Wright NM, Warren DK. The impact of prebiopsy antibiotics on pathogen recovery in hematogenous vertebral osteomyelitis. Clin Infect Dis. 2011;52:867–72. doi:10.1093/cid/cir062. Park KH, Chong YP, Kim SH, Lee SO, Choi SH, Lee MS, et al. Clinical characteristics and therapeutic outcomes of hematogenous vertebral osteomyelitis caused by methicillin-resistant Staphylococcus aureus. J Infect. 2013;67:556–64. doi:10.1016/j. jinf.2013.07.026. Bhavan KP, Marschall J, Olsen MA, Fraser VJ, Wright NM, Warren DK. The epidemiology of hematogenous vertebral osteomyelitis: a cohort study in a tertiary care hospital. BMC Infect Dis. 2010;10:158. doi:10.1186/1471-2334-10-158. Bauman GI, Stifel RE. Osteomyelitis of the spine. Ann Surg. 1923;78:119–21. Lillie P, Thaker H, Moss P, Baruah J, Cullen L, Taylor D, et al. Healthcare associated discitis in the era of antimicrobial resistance. J Clin Rheumatol. 2008;14:234–7. doi:10.1097/RHU. 0b013e318181addd.
