Management of vertebral osteomyelitis over an eight-year period: The UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis)

Journal Pre-proof Management of vertebral osteomyelitis in an 8-year period: the UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis) Alessandro Russo, Elena Graziano, Alessia Carnelutti, Massimo Sponza, Barbara Cadeo, Assunta Sartor, Elda Righi, Matteo Bassetti

PII:

S1201-9712(19)30403-5

DOI:

https://doi.org/10.1016/j.ijid.2019.10.010

Reference:

IJID 3786

To appear in:

International Journal of Infectious Diseases

Received Date:

13 August 2019

Revised Date:

7 October 2019

Accepted Date:

11 October 2019

Please cite this article as: Russo A, Graziano E, Carnelutti A, Sponza M, Cadeo B, Sartor A, Righi E, Bassetti M, Management of vertebral osteomyelitis in an 8-year period: the UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis), International Journal of Infectious Diseases (2019), doi: https://doi.org/10.1016/j.ijid.2019.10.010

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier.

Management of vertebral osteomyelitis in an 8-year period: the UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis)

Running head: management of vertebral osteomyelitis

Alessandro Russo1, Elena Graziano1, Alessia Carnelutti1, Massimo Sponza2, Barbara Cadeo1,

ro of

Assunta Sartor3, Elda Righi4, Matteo Bassetti1

1

re

-p

Infectious Diseases Clinic, Department of Medicine, University of Udine and Azienda Sanitaria Universitaria Integrata di Udine, Udine, Italy 2 Division of Vascular and Interventional Radiology, University of Udine and Azienda Sanitaria Universitaria Integrata di Udine, Udine, Italy 3 Microbiology Unit, Azienda Sanitaria Universitaria Integrata, Udine, Italy. 4 Infectious Diseases, Department of Diagnostics and Public Health, University of Verona, Verona, Italy.

Corresponding author: Matteo Bassetti, Clinica Malattie Infettive, Azienda Sanitaria Universitaria Integrata di Udine, Presidio Ospedaliero Universitario Santa Maria della

lP

Misericordia Piazzale Santa Maria della Misericordia 15, 33100 Udine, Italy, Phone +39 0432

Highlights Summary

na

559355; Fax +39 0432 559360, Email: [email protected]

Etiology of vertebral osteomyelitis is globally obtained in less than 40% of cases



UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis) was applied to all cases



Rigorous application of protocol permitted to obtain about 74% of etiology



Antimicrobial wash-out plus vertebral biopsy increased the rate of diagnosis



Use of FDG-PET/TC was more reliable in follow-up of infection than MRI

Jo

ur



ABSTRACT Objectives: vertebral osteomyelitis (VO) is a compelling clinical entity for clinicians, because of its insidious and indolent course which make diagnosis difficult.

1

Methods: all patients with suspected diagnosis of VO were analyzed over an 8-year period (January 2009 - January 2017). The UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis) protocol was applied to all cases. Primary endpoint was performance of UDIPROVE to obtain causal bacteria of infection. Results: during study period, 133 episodes of confirmed VO were observed. Etiology of infection was obtained in 73.6% of cases: Gram-positive were 70.5%, Gram-negative 16.3%, mycobacteria 13.2%. FDG-PET/TC showed that for tubercular VO median SUV was higher if

ro of

compared with other bacteria. Clinical cure at end of therapy was reported in 85.7% of patients. Previous antimicrobial therapy and >5 days delay in performing biopsy were

associated with undiagnosed etiology of VO. Targeted antibacterial therapy and follow-up

-p

with FDG-PET/TC were associated with clinical cure at end of therapy, while >2 vertebrae involved and inadequate drainage were associated with failure.

re

Conclusions: rigorous application of UDIPROVE protocol permitted to obtain causative

lP

pathogens of VO, about twice the data reported in literature. Use of FDG-PET/TC was more reliable in follow-up of infection if compared to MRI.

na

Keywords: vertebral osteomyelitis; vertebral biopsy; C-reactive protein; FDG-PET/TC; MRI

ur

INTRODUCTION

Jo

Vertebral osteomyelitis (VO) is a compelling clinical entity for clinicians, because of its insidious and indolent course which make diagnosis difficult. VO can often be delayed several months and may initially be misdiagnosed and mismanaged as a degenerative process. Consequently, patients often develop destructive lesions or neurological complications. In this setting, an elevated C-reactive protein (CRP) result in patients with back pain, though not specific, has a sensitivity that can range from 94% to 100% [1-2].

2

Etiology of infection is globally obtained in less than 40% of cases, making the choice of antimicrobial regimen and duration of therapy challenging [3]. With the exception of septic patients or patients with neurologic compromise, empiric antimicrobial therapy should be withheld, when possible, until a microbiologic diagnosis is confirmed. An image-guided aspiration biopsy is recommended in patients with suspected VO when a microbiologic diagnosis has not been established by blood cultures (BC) or serologic tests; however, rate of diagnosis with an image-guided biopsy in literature is about 30-40% [4].

ro of

No indications for a follow-up with magnetic resonance imaging (MRI) in patients with VO in whom a favorable clinical and laboratory response to antimicrobial therapy is

recommended by international guidelines [4], considering the limits of the methodic. For

-p

these reasons, in the last years the role of positron emission tomographic (PET) scanning was assessed, considering also its highly sensitive for detecting osteomyelitis. A negative PET

re

scan excludes the diagnosis of osteomyelitis, including VO [5-6].

lP

Aim of the study was rigorous application of UDIPROVE (UDIne PROtocol on VErtebral osteomyelitis) protocol to obtain causative pathogens of VO with vertebral biopsy, and the role of F18-fluorodeoxyglucose positron emission tomography/computed tomography (FDG-

na

PET/CT) in diagnosis and follow-up of infection and comparison with MRI.

ur

MATERIALS AND METHODS

Jo

Study design

We conducted a prospective observational study in in a large tertiary-hospital in Italy over a 8-year period (January 2009 - January 2017). During the study period, all consecutive adult patients (age ≥18 years) with suspected diagnosis of VO were analyzed. The UDIPROVE protocol was applied to all cases, as reported in Figure 1. The study was preapproved by the local ethics committee and conducted in accordance with the principles of the Declaration of Helsinki. 3

Data collection and Definitions Patient data from medical records, computerized hospital databases, and/or clinical charts were prospectively collected using a standard form. They included demographics, clinical and laboratory findings, baseline comorbidities, radiological findings, relapse of infection, microbiological data, the Charlson Comorbidity Index, date and type of surgery (if performed), source of infection, persistent bacteremia, development of septic shock, duration of hospital stay, duration of definitive antibiotic therapy, clinical treatment failure, and 30-day

ro of

mortality. Patients performed at time of diagnosis MRI and/or FDG-PET/CT (1st visit). All patients who performed a FDG-PET/CT were on a low carbohydrate-fat allowed diet 24 h before FDG-

-p

PET/CT was performed, and they fasted 6 h before 18F–FDG-injection, according with local protocol.

re

The following definitions were established prior to data analysis: relapse - a new diagnosis of

lP

VO caused by the same organism after clinical and microbiological resolution of a previous episode treated VO; time to biopsy - the number of days that elapsed between diagnosis of VO and vertebral biopsy (when performed); persistent infection - patients still treated for VO at

na

three month follow-up without resolution of the described symptoms; clinical treatment failure - lack of response to the definitive antimicrobial regimen, as reflected by the presence

ur

of any of the following after ≥ 45 days of therapy: ongoing fever, leukocytosis, or other

Jo

clinical signs of infection that could not be attributed to causes other than VO; 30-day mortality - death from any cause within the 30 days following diagnosis of VO. Synchronous or metastatic foci of infection were identified by means of clinical and imaging examinations. When present, fluid collections and abscesses were drained and the fluid cultured. Coagulase-negative staphylococci (CoNS) and other skin commensals were considered etiologically irrelevant unless they were isolated from two or more sets of BC and their causative role was consistent with clinical data. Adequate control of source of infection 4

was defined as the removal of any preexisting contaminated CVC as well as the drainage of vertebral abscesses or other fluid collections have been performed within 24 h after the onset of infection. Patients who were diagnosed with VO based on at least one imaging study and/or blood or tissue culture results were treated according to the IDSA (Infectious Diseases Society of America) guideline for vertebral osteomyelitis [4]. Three months after inclusion (2nd visit), all surviving patients performed MRI and/or FDG-PET/CT. Endpoints and Statistical Analysis

ro of

Primary endpoint was performance of UDIPROVE to obtain causal bacteria of infection; secondary endpoints were role of FDG-PET/CT in diagnosis and follow-up and analysis of clinical cure at end of therapy.

-p

Between-group differences were assessed with the chi-square test or Fisher exact test (for categorical variables) and the two-tailed t test or Mann-Whitney test (for continuous

re

variables), as appropriate. Univariate and multivariate analysis were used to determine the

lP

effects of different variables on undiagnosed etiology of VO, clinical treatment failure and clinical success in the study population. Possible confounding factors and interactions were weighted during analyses with a backward stepwise selection and considering p ≤ 0.05 for all

na

variables, to determine the effects of all anamnestic, clinical and therapeutic variables on outcomes. All reported P values are two-tailed, wald confidence intervals and odd ratios were

ur

computed based on estimated standard errors. All analyses were done with the SPSS software

Jo

package (version 20.0, SPSS Inc., Chicago, Illinois).

RESULTS

During the study period, 133 episodes of confirmed VO were observed. As reported in Table 1, etiology of infection was obtained in 98 (73.6%) cases (BC + vertebral biopsy/drainage) according to the UDIPROVE algorithm (Figure 1). Most frequently isolated bacteria were Gram-positive (70.5%), followed by Gram-negative (16.3%), and mycobacteria (13.2%). 5

Among Gram-positive, Streptococcus spp. accounted for 24.4% of cases, followed by MRSA (19.4%). Biopsy results with or without antimicrobial wash-out are reported in Figure 2. Mean SUV, CRP and WBC values at time of diagnosis (1st visit) and at follow-up (2nd visit) are reported in Figure 3. Statistically significant differences were observed for SUV (p<0.001) and CRP (p<0.001) values. Finally, in Figure 4 are reported SUV values in subgroups of patients with VO at follow-up (2nd visit): patients with or without tubercular etiology, presence or absence of endocarditis,

ro of

Staphylococcus aureus etiology or others, and 2 or >2 vertebrae involved in infection. Statistically significant differences in SUV values were observed for tubercular etiology or not (p<0.001) at follow-up.

-p

Table 2 reports baseline characteristics and outcome of the study population along with

univariate analysis of risk factors for clinical success or failure. Vertebral biopsy/drainage was

re

performed in 101 (75.9%) patients, and 14-day antimicrobial treatment discontinuation before

lP

vertebral biopsy was obtained in 98 (73.7%) patients. Concomitant endocarditis was observed in 18 (13.5%) of cases. Mean SUV value was 7.2 ± 3.3, and diagnosis and follow-up with FDG-PET/CT was performed in 59 (44.3%) patients. All patients performed MRI at time of

na

diagnosis and at follow-up. Clinical cure at end of therapy was reported in 114 (85.7%) of patients, and 30-day mortality was 10.5%.

ur

Univariate analysis of risk factors for clinical success or failure showed that >2 vertebrae

Jo

involved (94.7% Vs 24.5%, p<0.001), mean SUV value (7.6 ± 3.8 Vs 7 ± 3.6, p=0.03), and inadequate source control of infection (84.2% Vs 7.9%, p<0.001) were more frequently observed in patients with failure of therapy, compared with clinical cure group. Conversely, targeted antibacterial therapy (78.9% Vs 42.1%, p<0.001), and follow-up with FDG-PET/CT (50% Vs 10.5%, p<0.001) were more frequently reported in clinical cure group. No differences were observed in 30-day mortality. Follow-up with FDG-PET/TC showed that decrement of standardized uptake value (SUV) >30% from basal value was associated with 6

clinical cure at end of therapy. For tubercular VO, the mean SUV at time of diagnosis was higher if compared with other bacteria (11.4 ± 3.7 Vs 7 ± 2.3, p<0.001). Logistic regression analysis showed that previous antimicrobial therapy and >5 days delay in performing biopsy were associated with undiagnosed etiology of VO, as reported in Table 3A. Targeted antibacterial therapy and follow-up with FDG-PET/TC were associated with clinical cure at end of therapy, while >2 vertebrae involved and inadequate drainage of

ro of

infection were associated with failure of therapy (see Table 3-B).

DISCUSSION

Our data showed that rigorous application of UDIPROVE protocol permitted to obtain

-p

causative pathogens of VO in almost 75% of cases, a higher rate if compared with data

reported in literature [4]. Of importance, use of FDG-PET/TC was more reliable in follow-up

re

of infection, if compared to MRI. As consequence, targeted antibacterial therapy was

lP

independently associated with clinical cure of VO. Finally, previous antimicrobial therapy and >5 days delay in performing biopsy were associated with undiagnosed etiology of VO,

treatment failure.

na

while >2 vertebrae involved and inadequate drainage of infection were associated with

As matter of fact, the UDIPROVE protocol was associated with high rate of diagnosis in

ur

patients with VO. In literature, the sensitivity of the image-guided biopsy in evaluated studies

Jo

is about 40% [7]. Bhavan and coworkers showed that the yield of causative pathogens was found to increase with open biopsy (93%) compared with needle aspiration biopsy (48%) [8]. Of importance, VO is a very difficult to diagnose and treat infection, considering also the high rate of patients on antibiotic therapy at time of diagnosis, and this observation results in a low rate of positivity for BC and vertebral biopsy to determine etiology of infection [9]. As reported in Figure 2, also in our study vertebral biopsy without antimicrobial discontinuation, according with protocol, was associated with a high rate of negative results. 7

There are no studies to guide a clinician on managing patients with significant clinical or radiographic evidence of VO and who have sterile blood cultures and an initial non-diagnostic spinal aspiration biopsy result. Obtaining systemic inflammatory markers and diagnosis and follow-up with MRI may confirm infection and monitor response to therapy, clarify the presence of abscess in need of drainage, and identify spinal instability that could benefit from surgical correction. In previous studies, patients with evidence of progressive epidural and/or paraspinal soft tissue infection on follow-up MRI appear to be at a greater risk for treatment

ro of

failure [10]. The frequency and utility of obtaining follow-up inflammatory laboratory markers (like CRP) while patients are receiving antimicrobial therapy for VO have not been established. MRI has some value in assessing response of soft tissue infection, but bone and

-p

disk abnormalities typically appear to worsen even in successfully treated patients [11-12]. In our study the combination of FDG-PET/TC plus CRP values at time of diagnosis and follow-

re

up seems to be associated with higher accuracy to monitor response to therapy. FDG-PET/TC

lP

could be used in association with MRI for the detection of VO, but can also contribute to the early determination of the response to therapy. Our data are in line with previous attempts to assess treatment response that have relied on quantification of SUV value before and after

na

treatment in patients with VO [13-14].

Of interest, for tubercular VO the mean SUV was higher if compared with compared with

ur

pyogenic spondylodiscitis. Delayed diagnosis of tubercular etiology may lead to severe

Jo

consequences, including permanent spinal damage, and although spinal biopsies often represent the most accurate diagnostic methods, low microbiological yields can be associated with tubercular VO. In this setting, imaging therefore represents a key diagnostic tool to achieve early diagnosis of the disease and to monitor the response to antibiotic treatment. As previously observed [15], the association of specific FDG-PET/TC features, such as high SUV, with tubercular VO has been reported but remains to be definitely determined [16-17].

8

Then, the confirmation of high SUV values could help physicians to precociously hypothesize a tubercular etiology for VO [18]. However, in our study population, failure of therapy for VO was observed in 19 (14.3%) patients. The failure rates of treated patients with VO in most clinical studies varied between 10% and 30% [4- [19]. Therapeutic management of VO patients with treatment failure should be tailored to the suspected reason for failure. Consultation with a surgeon and infectious disease physician experienced in the treatment of spinal infections may be warranted in

ro of

patients with suspected or proven treatment failure. The decision of whether surgical intervention is warranted needs to be individualized, and incorporates similar principles as to whether to perform surgery at the time of VO diagnosis or not [20].

-p

Some important limitations of this study should be highlighted. First of all, there is not a preintervention period that reduces the quality of evidence, since this is a descriptive analysis of

re

patients managed according to the UDIPROVE protocol. Second, its single-center nature and

lP

the short follow-up period (3 months) reduces generalizability and comparability of these results. Finally, FDG-PET/TC was not performed in all study population and its routine use could be limited due to the high costs.

na

In conclusion, etiology of infection was obtained in 73.6% cases (combination of BC plus vertebral biopsy/drainage), according to the UDIPROVE algorithm; moreover, combination

ur

of protocol with use of FDG-PET/TC for diagnosis and follow-up was associated with high

Jo

rate of clinical cure (85.7%) in a very difficult-to-treat infection like VO. Finally, antimicrobial therapy at time of vertebral biopsy influenced negatively the possibility to obtain etiology of infection. The rigorous application of antimicrobial wash-out in combination with vertebral biopsy and collection of BC greatly increase the rate of diagnosis of VO causative pathogens. Finally, higher SUV values could be associated with tubercular etiology, and the combination of SUV plus CRP values could be helpful to monitor adequate response to therapy in all patients. 9

AKNOWLEDGMENTS Funding: none. Conflict of interests: none. Ethical approval: the study was preapproved by the local ethics committee. Author contributions: AR and MB designed the study: EG, AC, MS, BC, ER collected data;

Jo

ur

na

lP

re

-p

ro of

AR, ER, MB wrote the manuscript

10

REFERENCES 1

Zimmerli W. Clinical practice. Vertebral osteomyelitis. N Engl J Med 2010; 362:1022–9.

2

Gupta A, Kowalski TJ, Osmon DR, Enzler M, Steckelberg JM, Huddleston PM, et al. Long-term outcome of

pyogenic vertebral osteomyelitis: a cohort study of 260 patients. Open Forum Infect Dis 2014; 1:8. 3

Chew FS, Kline MJ. Diagnostic yield of CT-guided percutaneous aspiration procedures in suspected spontaneous

infectious diskitis. Radiology 2001; 218:211–4. 4

ro of

Berbari EF, Kanj SS, Kowalski TJ, Darouiche RO, Widmer AF, Schmitt SK, et al. 2015 Infectious Diseases

Society of America (IDSA) clinical practice guidelines for the diagnosis and treatment of native vertebral osteomyelitis in adults. Clin Infect Dis 2015;61:e26–e46 5

-p

Ohtori S, Suzuki M, Koshi T, Yamashita M, Yamauchi K, Inoue G, et al. 18F-fluorodeoxyglucose-PET for

patients with suspected spondylitis showing Modic change. Spine 2010; 35:E1599–603. 6

re

Hungenbach S, Delank KS, Dietlein M, Eysel P, Drzezga A, Schmidt MC. 18F-fluorodeoxyglucose uptake

7

lP

pattern in patients with suspected spondylodiscitis. Nucl Med Commun 2013;34:1068–1074 Sehn JK, Gilula LA. Percutaneous needle biopsy in diagnosis and identification of causative organisms in cases

8

na

of suspected vertebral osteomyelitis. Eur J Radiol 2012; 81:940–6. Bhavan KP, Marschall J, Olsen MA, Fraser VJ, Wright NM, Warren DK. The epidemiology of hematogenous

9

ur

vertebral osteomyelitis: a cohort study in a tertiary care hospital. BMC Infect Dis 2010; 10:158. Saeed K, Esposito S, Ascione T, Bassetti M, Bonnet E, Carnelutti A, et al; ISAC Skin, Bone and Joint Infection

Jo

Working Group. Hot Topics on Vertebral Osteomyelitis from the International Society of Antimicrobial Chemotherapy.

Int

J

Antimicrob

Agents.

2019

Jun

13.

pii:

S0924-8579(19)30166-9.

doi:

10.1016/j.ijantimicag.2019.06.013. [Epub ahead of print] 10

Kowalski TJ, Layton KF, Berbari EF, Steckelberg JM, Huddleston PM, Wald JT, et al. Follow-up MR imaging

in patients with pyogenic spine infections: lack of correlation with clinical features. AJNR Am J Neuroradiol 2007; 28:693–9

11

11

Zarrouk V, Feydy A, Sallès F, Dufour V, Guigui P, Redondo A, et al. Imaging does not predict the clinical

outcome of bacterial vertebral osteomyelitis. Rheumatology (Oxford, England) 2007;46(2):292–295 12

Carragee E. The clinical use of magnetic resonance imaging in pyogenic vertebral osteomyelitis. Spine (Phila

Pa 1976) 1997;22(7):780–785 13

Nanni C, Boriani L, Salvadori C, Zamparini E, Rorato G, Ambrosini V, et al. FDG PET/CT is useful for the

interim evaluation of response to therapy in patients affected by haematogenous spondylodiscitis. Eur J Nucl Med Mol Imaging 2012;39(10):1538–1544 Kim S-JSJ, Kim I-J, Suh KT, Kim Y-K, Lee JS. Prediction of residual disease of spine infection using F-18

FDG PET/CT. Spine (Philadelphia, Pa 1976) 2009;34:2424–2430 23 15

ro of

14

Bassetti M, Merelli M, Di Gregorio F, Della Siega P, Screm M, Scarparo C, et al. Higher fluorine-18

spondylodiscitis. Skeletal Radiol. 2017 Jun;46:777-783.

-p

fluorodeoxyglucose positron emission tomography (FDG-PET) uptake in tuberculous compared to bacterial

Bakheet SM, Powe J, Ezzat A, Rostom A. F-18-FDG uptake in tuberculosis. Clin Nucl Med. 1998;23:739–42.

17

Trecarichi EM, Di Meco E, Mazzotta V, Fantoni M. Tuberculous spondylodiscitis: epidemiology, clinical

re

16

18

lP

features, treatment, and outcome. Eur Rev Med Pharmacol Sci. 2012;16 Suppl 2:58–72. Heysell SK, Thomas TA, Sifri CD, Rehm PK, Houpt ER. 18-fluorodeoxyglucose positron emission tomography

19

Rissing JP. Antimicrobial therapy for chronic osteomyelitis in adults: role of the quinolones. Clin Infect Dis

ur

1997; 25:1327–33.

Valancius K, Hansen ES, Hoy K, Helmig P, Niedermann B, Bunger C. Failure modes in conservative and

Jo

20

na

for tuberculosis diagnosis and management: a case series. BMC Pulm Med. 2013;13:14.

surgical management of infectious spondylodiscitis. Eur Spine J 2013; 22:1837–44.

12

ro of

Figure 1. UDIPROVE algorithm

-p

Legend. ID: infectious diseases; TB: tuberculosis.

Figure 2. Biopsy results with or without antimicrobial wash-out

re

90 80

lP

70 60 50 40

na

30 20

0

ur

10

Jo

Biopsy during antimicrobial therapy

Positive

Biopsy after antimicrobial wash-out

Negative

Figure 3. Mean SUV, CRP and WBC values at time of diagnosis (1st visit) and at follow-up (2nd visit)

13

ro of

Legend. SUV: standard uptake value; CRP: c-reactive protein; WBC: white blood cells.

Jo

ur

na

lP

re

-p

Figure 4. SUV values in subgroups of patients with VO at follow-up (2nd visit)

Legend. SUV: standard uptake value; VO: vertebral osteomyelitis; TB: tuberculosis.

14

N=98 (%)

Gram-positive

69 (70.5)

MRSA

19 (19.4)

MSSA

5 (5.1)

CoNS

13 (13.3)

Streptococcus spp

24 (24.4)

Enterococcus spp

4 (4.1)

Others

4 (4.1)

Gram-negative

16 (16.3)

Escherichia coli

8 (8.1)

Klebsiella pneumoniae

4 (4.1)

Others

4 (4.1)

Mycobacterium tuberculosis

13 (13.3)

-p

Etiology

ro of

Table 1. Etiology of VO

re

Legend. VO: vertebral osteomyelitis; MRSA: methicillin-resistant Staphylococcus

aureus; MSSA: methicillin-sensitive Staphylococcus aureus; CoNS: coagulase-negative

Jo

ur

na

lP

staphylococci.

15

Table 2. Baseline characteristics and outcome of study population and univariate analysis of risk factors for clinical cure or failure of therapy All population Clinical cure n= 133 (%) n= 114 (%)

Male sex

65.7 ± 13.9 68 (51.1)

Charlson Comorbidity Index, mean ± SD

2.5 ± 1.4

Previous spinal surgery

20 (15) 112 (84.2)

Age, mean ± SD

Previous antimicrobial therapy (30 days) 14-day antimicrobial treatment discontinuation before vertebral biopsy >2 vertebrae involved

98 (73.7) 46 (34.5) 18 (13.5)

Other localization of infection

41 (30.8)

Biopsy/drainage performed

101 (75.9)

Blood culture positivity

34 (25.5)

SUV value, mean ± SD CRP concentration (mg/L), mean ± SD

7.2 ± 3.3 42.2 ± 32.9 8.3 ± 5.7 25 (18.8) 11.6 ± 7.8 98 (73.6) 10.9 ± 8.9 41.9 ± 26.3 34 ± 17.2 59 (44.3) 14 (10.5)

na

lP

Inadequate source control of infection Time to biopsy, mean ± SD Targeted antibacterial therapy Time to definitive therapy, mean ± SD Length of definitive therapy, mean ± SD

re

WBC concentration (109/L), mean ± SD

Length of hospitalization, mean ± SD Use of FDG-PET/CT 30-day mortality

58 (50.8) 2.4 ± 1.5 18 (15.7) 95 (83.3)

2.5 ± 1.6

1.0

2 (10.5) 17 (89.4)

0.73 0.74

0.94 1.0

85 (74.5)

13 (68.4)

0.58

28 (24.5)

18 (94.7)

<0.001

15 (13.1)

3 (15.7)

0.91

30 (26.3)

11 (57.8)

0.01

88 (77.1)

13 (68.4)

0.34

30 (26.3) 7 ± 3.6 40.2 ± 31.7 8.4 ± 5.2

4 (21)

0.62

7.6 ± 3.8 44.9 ± 33.9 8.1 ± 5

0.03 0.04 0.81

9 (7.9) 10.5 ± 8.2 90 (78.9) 11.2 ± 9.2 43.2 ± 27 33.6 ± 15.1

16 (84.2) 12.9 ± 9.2 8 (42.1) 10.2 ± 8.7 40.1 ± 24.9 34.2 ± 20.5 2 (10.5) 2 (10.5)

<0.001 0.03 <0.001 0.81 0.07 0.92

-p

Endocarditis

66.3 ± 14.1

p

ro of

VARIABLES

Failure of therapy n=19 (%) 65.6 ± 13.4 10 (52.6)

57 (50) 12 (10.5)

<0.001 1.0

ur

Legend. SUV: standard uptake value; CRP: c-reactive protein; WBC: white blood cells; FDG-PET/CT: F18-

Jo

fluorodeoxyglucose positron emission tomography/computed tomography.

16

Table 3. Variables independently associated with undiagnosed etiology of VO (A), clinical cure and failure of therapy (B) A

B

Variables

OR

CI95%

P

Previous

2.3

1.98-

<0.001 Targeted

antimicrobial

Variables

3.12

OR

CI95%

P

0.34

0.21-

0.01

antibacterial

therapy

0.54

therapy 4.2

performing biopsy

3.12-

<0.001 Follow-up with

7.23

0.45

FDG-PET/CT >2 vertebrae

-p

involved Inadequate

re

drainage of

0.31-

0.02

ro of

>5 days delay in

0.78

2.87

1.89-4.2

<0.001

4.21

2.87-

<0.001

8.23

lP

infection

Legend. VO: vertebral osteomyelitis; FDG-PET/CT: F18-fluorodeoxyglucose positron emission

Jo

ur

na

tomography/computed tomography.

17