- Email: [email protected]
Surgical site infection in spinal metastasis: incidence and risk factors
Accepted Manuscript Title: Surgical site infection in spinal metastasis: incidence and risk factors Author: Amer Sebaaly, Daniel Shedid, Ghassan Boubez, Fahed Zairi, Michelle Kanhonou, Sung-Joo Yuh, Zhi Wang PII: DOI: Reference:
S1529-9430(18)30004-4 https://doi.org/10.1016/j.spinee.2018.01.002 SPINEE 57566
To appear in:
The Spine Journal
Received date: Revised date: Accepted date:
19-9-2017 12-12-2017 10-1-2018
Please cite this article as: Amer Sebaaly, Daniel Shedid, Ghassan Boubez, Fahed Zairi, Michelle Kanhonou, Sung-Joo Yuh, Zhi Wang, Surgical site infection in spinal metastasis: incidence and risk factors, The Spine Journal (2018), https://doi.org/10.1016/j.spinee.2018.01.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.
1
Surgical site infection in spinal metastasis: incidence
2
and risk factors
3 4
Authors:
5
Amer SEBAALY (1, 2), Daniel SHEDID (3), Ghassan BOUBEZ (1), Fahed ZAIRI (3, 4),
6
Michelle KANHONOU (5), Sung-Joo YUH (3), Zhi WANG (1)
7 8 9 10
Affiliations: 1. Department of Orthopedic surgery, Spine Unit, Centre Hopitalier de l’Université de Montréal (CHUM), Montréal, Canada
11
2. Saint Joseph University, Faculty of medicine, Beirut, Lebanon
12
3. Department of Neurosurgery, Spine Unit, Centre Hopitalier de l’Université de Montréal
13
(CHUM), Montréal, Canada
14
4. Department of neurosurgery, Lille university hospital, Lille France
15
5. Research center of Montreal University, CR-CHUM, Montreal, Canada
16
Corresponding author:
17
Amer SEBAALY
18 19
Department of Orthopedic surgery, Spine Unit, Centre Hopitalier de l’Université de Montréal (CHUM), Montréal, Canada
20
Saint Joseph University, Faculty of medicine, Beirut, LEBANON
21
Hôpital Notre Dame, Rue Sherbrooke, Montréal, QC, H2L 4M1, Canada
22
Email: [email protected]
Tel +961 3 679 128 1 Page 1 of 22
1
Abstract:
2
Background: Surgical site infection (SSI) inspinal metastasis surgery represents the most
3
common postoperative surgical complication with high morbidity and mortality.
4
Objective: Evaluate the incidence of SSI in spinal metastasis surgery and its risk factors.
5
Study Design: Retrospective analysis of a prospective collected data.
6
Methods: Preoperative, operative and post-operative data were collected together with the
7
Modified Tokuhashi score and Frankel score at all time checkpoints. SSI was divided into
8
superficial and deep SSI as well as early (<90 days) and late SSI. Multiple logistic regression
9
analysis was performed to identify independent risk factors with p<0.05 as significance
10
threshold.
11
Results: 297 patients were included with an incidence of SSI of 5.1% (Superficial SSI:
12
3.4%; deep SSI: 1.7 %). Cervicothoracic surgery was associated with the highest incidence of
13
SSI while cervical surgery had the lowest incidence. Smoking, higher number of spinal
14
metastasis, elevated BMI and higher ASA score were the preoperative factors associated with
15
increased risk of SSI. Increased intraoperative blood loss and increased number of fixed vertebra
16
increased the SSI incidence. SSI increased hospital stay by a mean of 12 days. When all these
17
variables are analyzed in a multiple regression model, only surgical time ≥ 4 hours and ASA ≥ 3
18
were found to be independent risk factor for the occurrence of SSI.
19
Conclusion: This paper represents the largest series of spinal metastasis with a mean
20
incidence of SSI of 5.1%. Smoking, higher BMI, higher number of spinal metastasis, higher
21
ASA score, higher number of fused vertebra, intraoperative bleeding ≥ 2000 ml and neurological
22
deterioration are risk factors for SSI occurrence. Only ASA ≥ 3 and operative duration ≥ 4 hours 2 Page 2 of 22
1
are independent risk factors for this complication occurrence. Finally, SSI occurrence is
2
associated with increased hospital stay, increased 30 days mortality rate and decreased survival
3
rates.
4 5
Level of evidence: Level IV
6 7 8
Keywords: Spinal metastasis; Surgery; fixation; surgical site infection; acute infection; debridement.
3 Page 3 of 22
1
Introduction:
2
The axial skeleton is the third most frequent location of metastasis (after the liver and the
3
lungs) and the most common site of bone metastasis [1,2]. The incidence of spinal metastasis is
4
increasing with the improvement of treatment of primary tumors as well as the improvement of
5
palliative therapy [3]. Postmortem studies have shown that the incidence of spinal metastasis
6
ranges between 30-90% depending on the primary cancer type [4]. Ideal management of spinal
7
metastasis requires a multidisciplinary approach that involves radiation therapy (conventional /
8
Stereotactic Body Radiation Therapy (SBRT)), chemotherapy, immunotherapy (targeted
9
therapy), cement augmentation (vertebroplasty, kyphoplasty) and surgery (en bloc resection,
10
debulking or palliative surgery) [3,5]. The later is indicated in case of rapidly progressive
11
neurological deficit from neurological compression of the spinal cord by the tumor [6,7], the
12
need to establish diagnosis, radioresistant tumors, neurological deterioration under radiotherapy
13
or in the case of impending instability based in the Spinal Instability Neoplastic Score (SINS)
14
[2,8].
15
Even though recent reports have showed excellent functional outcomes [9], perioperative
16
complications of this spinal surgery are relatively high and range from 5.3 to 76.2% with an
17
average of 26.9% [3]. Surgical site infection (SSI) represents the most common postoperative
18
surgical complication [3,9]. The exact incidence of spinal SSI following metastasis surgery is
19
heterogenous in the literature and ranges between 3.51% and 20% with a mean of 10.22% [5,10–
20
25]. This incidence is eight to ten times higher than the mean SSI rate of 2.2% of spinal surgery
21
in general [1,9,26]. Even more, some authors found that spinal oncologic surgery is an
22
independent factor for SSI (OR=6.2, CI=1.7-22.3) [27]. Identically, the risk of mortality
23
increases by two folds with patients with SSI and by 1.5 fold for a stay in intensive care unit 4 Page 4 of 22
1
[28]. Several risk factors have been proposed for SSI: advanced age, low albumin levels,
2
multiple levels of surgery, use of preoperative steroids, closure with non resorbable skin suture,
3
surgery duration ≥ 4 hours, bleeding ≥ 3000 ml, preoperative neurological compromise, number
4
of instrumented vertebras, diabetes and preoperative radiation therapy [3,9,21,29]. Nonetheless,
5
these conclusions emerge from old retrospective studies with inhomogeneous inclusion criteria
6
and small sample sizes. Even more, these studies were not powered enough to conduct multiple
7
regression analysis to detect independent risk factor for SSI occurrence.
8
At our institution, a tertiary spine referral center, spinal metastasis surgery is one of the
9
most undergone surgeries (between 40-60 cases a year). Thus, since January 2010, a prospective
10
database was created to include all cases of spinal tumor surgery (primary and metastasis). The
11
objectives of this study are to analyze the incidence of SSI in this large database while trying to
12
analyze possible dependent and independent risk factors SSI as well as the effect of SSI on
13
hospital stay and overall mortality.
14 15
Material and methods:
16
This is a retrospective study of a prospective collected data from the spinal metastasis
17
surgery monocentric database. This study was approved by the institutional ethical board.
18
Inclusion criteria included age greater than 18 years old and patients operated for a spinal
19
metastasis of a known or unknown primary tumor. The main indication for operation was
20
epidural compression form the tumor and/or instability. Instability was assessed by the presence
21
of a burst fracture or severe osteolysis involving the anterior and posterior elements based on the
22
SINS score [8]. The data collection started on September 2011 (start of using the SINS criteria)
5 Page 5 of 22
1
till the end of December 2016. Patients were included consecutively and were operated by one of
2
the three senior surgeons (GB, DS, and ZW). Exclusion criteria were primary tumors (bone and
3
neurologic tumors), nonmalignant causes of spinal compression (benign tumors), as well as
4
patients whose neurological deficit was explained by another process (previous stroke). Patients
5
received postoperative adjuvant radiation therapy if deemed necessary by the radiation oncology
6
department. All patients received our prophylactic antibiotic protocol with cephazolin 1 to 2 g IV
7
30 min before incision repeated at 4 hours for lengthy procedures and cephazolin 1-2 g every 8
8
hours for 24h postoperatively (3 doses). If the patient has an allergy to penicillin, Vancomycin
9
replaces cephazolin. All patients received minimal 1L of irrigation before closure for these
10
complex procedures. Preoperative radiation therapy did not have a standardized protocol.
11
Patients had surgery within six months of the radiation therapy. As for postoperative radiation
12
therapy, the majority of the patients had a standard protocol depending on the primary tumor. All
13
patients received SBRT radiation with varying dose depending on the primary between 2 and 4
14
weeks postoperatively.
15
Data collection encompassed demographic data that included age, sex, and smoking
16
status. Tumor type, localization of the tumor as well as known primary status were also noted.
17
Preoperative data included Frankel score, modified Tokuhashi score, site(s) of compression,
18
number of spinal metastasis, history of radiation therapy, urinary sphincter function, preoperative
19
embolization of the spinal metastasis and preoperative hemoglobin level. Operative data included
20
type of procedure, bleeding volumes, ASA (American Society of Anesthesiologist) score, off-
21
hours surgery status (surgery starting between 17:00 and 8:00 the following day) as well as
22
surgery duration. Postoperative data is collected in the immediate postoperative period and at the
23
final follow-up and included postoperative complications (SSI, dural tear, pneumonia, delirium,
6 Page 6 of 22
1
cardiovascular complications including stroke, pulmonary embolism, others…), duration of stay,
2
last follow-up (FU) Frankel score, postoperative radiation and dose of the radiation therapy.
3
Survival analysis included survival rate and survival duration after surgery. This was last updated
4
on the start of the preparation of this manuscript (minimum 9 months follow-up for the last case).
5
Superficial SSI and deep SSI were defined according the CDC criteria [30]. Superficial
6
SSI is an infection involving only the dermis and the subcutaneous tissue without invading the
7
paraspinous fascia. Deep SSI is invades the fascia and is in direct contact with the implants and
8
is therefore more severe. When the patients presented with fever, erythema and tenderness, or
9
elevated CRP, high index of suspicion of deep SSI warranted MRI examination. In the absence
10
of clinical and imaging signs of invasion of the muscle fascia, the diagnosis of superficial SSI is
11
retained. Acute SSI was defined as SSI occurring during the first 90 days postoperatively while
12
late infection was defined as the one occurring after the 90 days of surgery. Patients were then
13
divided into two groups according to SSI incidence: patients in Group 1 had SSI, while patients
14
in group 2 didn’t (could have another complication). Moreover, some variables were
15
retrospectively collected for group 1 patients: number of Irrigation & Débridement (I&D),
16
antibiotherapy (duration and type of administration), removal of the instrumentation, and use of
17
vacuum assisted closure (VAC).
18
Statistical analysis is done using SPSS for Windows, Version 18.0. p values below 0.05
19
were considered statistically significant. Student t test was used to compare means. χ2 test was
20
used to compare binary variables and to calculate odds ratio. MANOVA test was used to identify
21
possible risk factors for SSI occurrence and finally a multiple logistic regression analysis was
22
made to identify independent risk factors.
23 7 Page 7 of 22
1
Results:
2
The database of primary and secondary tumors consisted of 458 cases and 297 patients
3
were included in this study since they were operated on for spinal metastasis between September
4
2011 and December 2016. Missing data represented less than 10% of the entire data.
5
Complication data was complete for all the cases. Incomplete data was considered missing
6
completely at random and thus ignored in statistical analysis. In this cohort, surgical indication
7
was neurologic compromise and/or instability according to the SINS criteria. The majority of
8
patients (90.5%) had posterior approach with decompression and fusion and 67% had
9
corpectomy. The remaining 9.5 % had anterior first corpectomy followed by posterior approach
10
for fixation. No case needed a flap for primary coverage. Demographic, preoperative, operative
11
and postoperative data are summarized in table 1. Around 8.7% of patients had neurologic
12
deterioration following surgery whereas 68.2% had no change in their Frankel score and 23.2%
13
had an improvement of minimum one Frankel score.
14
Mean SSI rate was 5.1% (15 patients). Superficial SSI had and incidence of 3.4% (10
15
patients) while deep SSI had an incidence of 1.7 % (number of cases = 5). Patients with deep SSI
16
were debrided by a mean of 1.2 debridements one patient requiring two debridements). All but
17
one patient presented with acute or subacute infection (mean time to infection 26.7 days) with
18
only one patient presenting with infection 8 months post-operatively. Staphylococcus aureus was
19
incriminated in the majority of cases (73%). One infection was caused by more than one
20
organism (Echerichia coli). Vancomycin was the most used IV antibiotic with a mean duration
21
of IV antibiotic was 16.2 days. One patient (delayed infection case) required instrumentation
22
removal and VAC closure, followed by instrumentation afterwards and flap coverage for closure.
23
There was a tendency for a seasonal effect with the lowest infection rate (2.6%) recorded in 8 Page 8 of 22
1
summer (July-September) whereas highest incidence rate was recorded in winter (January-
2
March; 8%; p=0.13). Cervicothoracic surgery was associated with the highest incidence of SSI
3
(12.5%) while cervical surgery had the lowest incidence (2.22%), with a tendency to statistical
4
significance when compared to dorsal surgery (SSI incidence= 5.92%; p=0.15). Lumbar surgery
5
has an SSI incidence of 2.94%.
6
When comparing patients with SSI with patients with no SSI (group 1 vs group 2), higher
7
smoking incidence was seen in group 1 (57.2% vs 23%; p=0.04). Mean number of involved
8
vertebrae was increased in group 1 (2.4) compared to group 2 (1.6 (p=0.03)). There was no
9
difference in mean Tokushashi score and no effect of preoperative diabetes status. Higher BMI
10
was noted in the group 1 (29.3 vs 24.5; p=0.07). Fifty seven percent of patients with SSI had a
11
history of radiation therapy compared to only 33% of patients with no SSI (p=0.009). Likewise,
12
ASA score was higher in group 1 (3) compared to group 2 (2.76; p=0.04) as well as for operative
13
bleeding (mean group 1= 2126 ml; mean group 2: 1224 ml; p=0.005). The SSI group had a
14
higher number of fixed vertebra (7.8 vs 5.5; p=0.03). SSI increased hospital stay by a mean of 12
15
days (25 in group 1 compared to 13 in group 2; p=0.002). Postoperative Ambulatory status did
16
not affect the occurrence of SSI (Non ambulatory: SSI rate of 6.1%; ambulatory: 3.92%; p =
17
0.4). Finally, mean survival with patients in group 1 was 120 days compared to 202 days in
18
group 2 (p=0.02).
19
Odds ratios were then calculated for the dependent variables. Smoking increased the risk
20
of SSI by 2.4 (1.06-7.2; p=0.04), ASA score ≥ 3 increased the odds of SSI by 1.1 (1.03-1.2;
21
p=0.02) while operative bleeding ≥ 2000 ml increased SSI by 2.66 (1.1-7; p=0.04). Neurological
22
deterioration was associated with an increased risk of SSI (OR= 4.2 (1.4-12.8; p=0.02). On the
23
other hand, the occurrence of any other complication (aside from SSI) did not increase the risk of 9 Page 9 of 22
1
SSI (OR=1.5 (0.5-3.9; p=0.5). Finally, surgical duration more than 4 hours was associated with
2
an increased risk of SSI (OR=3.76 (1.2-11.6; p=0.015). When all these variables are analyzed in
3
a multiple regression model, only duration ≥ 4 hours and ASA ≥ 3 were independent risk factor
4
for the occurrence of SSI.
5 6
Discussion:
7
SSI is a serious complication following spinal surgery in general and metastasis surgery
8
in particular. This study reports the incidence of SSI in a homogenous monocentric population of
9
patients with spinal metastasis surgery. This cohort represents the largest cohort in the literature
10
evaluating this complication with nearly 1.5 times the number of the largest reported series [5]).
11
The incidence of SSI was 5.1%, approximately half of the reported mean of 10.22% in the
12
literature [3]. Moreover, this is the first paper to study the effect of primary tumor type and
13
location on the occurrence of SSI with no effect of tumor type and a tendency of higher SSI in
14
dorsal spine compared to cervical spine [3]. Smoking, higher number of spinal metastasis, higher
15
ASA score, intraoperative bleeding ≥ 2000 ml and neurological deterioration were all found to
16
be risk factors for SSI occurrence. Only ASA ≥ 3 and operative duration ≥ 4 hours are
17
independent risk factors for this complication occurrence. Finally, SSI occurrence is associated
18
with increased hospital stay, increased 30 days mortality rate and decreased survival rates.
19
The incidence of SSI in this study is one of the lowest in the modern spine literature with
20
a 5.1% incidence of SSI (1.7% for deep SSI and 3.4% for superficial SSI). Atkinson et al found a
21
similar rate of deep (2%) with a higher superficial SSI (9.2%) with superficial to deep ratio of
22
4.6/1 [1]. In a similar manner, Demura et al found higher deep SSI incidence of 7.1% [9] and
10 Page 10 of 22
1
other studies found higher rate of deep SSI requiring reoperation (9.45%) [2]. This decreasing
2
rate of SSI along the years could be attributed to better infection prevention protocols and better
3
postoperative rehabilitations (Table 2; Figure 1). Reoperation rate for SSI was 2%. This finding
4
is much less compared to other studies in the literature (4.4 % reoperation rate in the study of
5
Quraishi et al [2]). Even more, I&D is needed in about one half of all SSI cases and all cases of
6
deep SSI, findings similar to the available literature [1,9]. This increased need for reoperation
7
exposes this particular population of frail patients to risk of recurrent anesthesia as well as other
8
surgical complications.
9
Many studies have evaluated the possible risk factors for the development of SSI.
10
Smoking is a known risk factor for SSI in general surgery and general spinal surgery but no
11
study has been able to show its correlation to SSI in spinal metastasis [25,31]. This is the first
12
study to find that preoperative smoking is a risk factor for SSI (OR= 2.4). On the other hand,
13
patients’ comorbidities are known risk factor for spinal SSI and constitute a category in the
14
PITSS scoring system for the need of I&D in deep spinal SSI [32]. Higher ASA scores ≥ 3,
15
denoting the presence of more co-morbidities, was found to be one important risk factor for the
16
occurrence of this complication. Even more, a lengthier surgery with higher bleeding volumes
17
are known risk factors for SSI as shown in this study [21,33]. Albeit neurologic deficits are
18
associated with higher risk of SSI in deformity or trauma surgery [26], there was no association
19
between ambulatory status and SSI as found by other authors in the literature [9,29]. Only
20
neurological deterioration after surgery was associated with increased SSI rates (OR=4.2).
21
Finally, the most controversial risk factor for SSI is preoperative radiation therapy. Classically,
22
this neoadjuvant therapy has negative effects on wound repair due to decreased vascularity,
23
hypoxia and impairment of tissue repair [3,9]. Some authors related this complication to the dose
11 Page 11 of 22
1
of radiation with increased complications with doses greater than 40 Gy [24]. Others found that
2
preoperative radiotherapy was associated with higher risk of deep SSI with an OR ranging from
3
3 to 41 [9,22,23]. Nonetheless, no increase of wound complications and SSI with the use of
4
radiation therapy was found in the study of Keam et al [31]. Similarly, we have not found
5
preoperative radiation therapy to be a risk factor of SSI.
6
The consequences of SSI in spinal metastasis surgeries are also debated. Duration of
7
hospitalization with patients with SSI increases by 100% (12 days) and is around 4 weeks (25
8
days). This finding is similar to the findings of other authors with a mean hospitalization duration
9
of 32 days [1]. To add to that, SSI affect survival after spinal metastasis surgery. We found a
10
decrease of survival by 4 months; same results were found by other authors with a one year
11
survivorship of 42% [34]. The study of Atkinson et al. didn’t have the statistical power to
12
confirm the impact of SSI on mortality. This decrease of survival could be associated with the
13
occurrence of SSI per se or the increased frailty of the patients who develop SSI: higher ASA
14
and lower Modified Tokuhashi scores. In a recent study, authors found that 38 to 75% of deaths
15
in patients with SSI are attributable to SSI itself [35]. Due to the small number of cases, the true
16
effect of deep versus superficial SSI on mortality could not be studied in this series and probably
17
needs multicentric efforts to increase the population number.
18
This study has several limitations. First of all, this is a prospectively collected database
19
that could have some missing data. Data collection, even thorough designed to be exhaustive,
20
could be imperfect. We considered missing data to be totally at random. Second of all, the low
21
rate of SSI could prevent us from detecting correlations and statistical difference on some
22
variables. Multicentric data pooling could help to increase the included number of patients and
12 Page 12 of 22
1
overcome this limitation. Nonetheless, this study represents the largest monocentric series in the
2
modern literature and drawn conclusions have solid clinical and statistical power.
3 4
Conclusion:
5
The incidence of SSI in the largest population of patients with spinal metastasis surgery
6
reaches 5.1%. Primary tumor type and location have no effect on SSI incidence. Smoking, higher
7
BMI, higher number of spinal metastasis, higher ASA score, higher number of fused vertebra,
8
intraoperative bleeding ≥ 2000 ml and neurological deterioration are risk factors for SSI
9
occurrence. Only ASA ≥ 3 and operative duration ≥ 4 hours are independent risk factors for this
10
complication occurrence. Finally, SSI occurrence is associated with increased hospital stay,
11
increased 30 days mortality rate and decreased survival rates.
13 Page 13 of 22
1 2
References: [1]
Atkinson RA, Jones A, Ousey K, Stephenson J. Management and cost of surgical site
3
infection in patients undergoing surgery for spinal metastasis. J Hosp Infect 2017;95:148–
4
53. doi:10.1016/j.jhin.2016.11.016.
5
[2]
Quraishi N a., Rajabian A, Spencer A, Arealis G, Mehdian H, Boszczyk BM, et al.
6
Reoperation rates in the surgical treatment of spinal metastases. Spine J 2015;15:S37–43.
7
doi:10.1016/j.spinee.2015.01.005.
8
[3]
9
DB. Perioperative Complications of Spinal Metastases Surgery. Clin Spine Surg
10 11
2017;30:4–13. doi:10.1097/BSD.0000000000000484. [4]
12 13
Luksanapruksa P, Buchowski JM, Zebala LP, Kepler CK, Singhatanadgige W, Bumpass
Wong DA, Fornasier VL, MacNab I. Spinal metastases: the obvious, the occult, and the impostors. Spine (Phila Pa 1976) 1990;15:1–4.
[5]
Lee BH, Park J-O, Kim H-S, Park Y-C, Lee H-M, Moon S-H. Perioperative complication
14
and surgical outcome in patients with spine metastases: retrospective 200-case series in a
15
single institute. Clin Neurol Neurosurg 2014;122:80–6.
16
doi:10.1016/j.clineuro.2014.04.025.
17
[6]
Dodwad S-NM, Savage J, Scharschmidt TJ, Patel A. Evaluation and treatment of spinal
18
metastatic disease. Cancer Treat Res 2014;162:131–50. doi:10.1007/978-3-319-07323-
19
1_7.
20 21
[7]
Patchell RA, Tibbs PA, Regine WF, Payne R, Saris S, Kryscio RJ, et al. Direct decompressive surgical resection in the treatment of spinal cord compression caused by
14 Page 14 of 22
1
metastatic cancer: a randomised trial. Lancet (London, England) 2005;366:643–8.
2
doi:10.1016/S0140-6736(05)66954-1.
3
[8]
Fisher CG, DiPaola CP, Ryken TC, Bilsky MH, Shaffrey CI, Berven SH, et al. A novel
4
classification system for spinal instability in neoplastic disease: an evidence-based
5
approach and expert consensus from the Spine Oncology Study Group. Spine (Phila Pa
6
1976) 2010;35:E1221-9. doi:10.1097/BRS.0b013e3181e16ae2.
7
[9]
Demura S, Kawahara N, Murakami H, Nambu K, Kato S, Yoshioka K, et al. Surgical Site
8
Infection in Spinal Metastasis. Spine (Phila Pa 1976) 2009;34:635–9.
9
doi:10.1097/BRS.0b013e31819712ca.
10
[10] Park S-J, Lee C-S, Chung S-S. Surgical results of metastatic spinal cord compression
11
(MSCC) from non-small cell lung cancer (NSCLC): analysis of functional outcome,
12
survival time, and complication. Spine J 2016;16:322–8.
13
doi:10.1016/j.spinee.2015.11.005.
14
[11] de Ruiter GCW, Lobatto DJ, Wolfs JF, Peul WC, Arts MP. Reconstruction with
15
expandable cages after single- and multilevel corpectomies for spinal metastases: a
16
prospective case series of 60 patients. Spine J 2014;14:2085–93.
17
doi:10.1016/j.spinee.2013.12.029.
18
[12] Bollen L, de Ruiter GCW, Pondaag W, Arts MP, Fiocco M, Hazen TJT, et al. Risk factors
19
for survival of 106 surgically treated patients with symptomatic spinal epidural
20
metastases. Eur Spine J 2013;22:1408–16. doi:10.1007/s00586-013-2726-4.
21
[13] Walter J, Reichart R, Waschke A, Kalff R, Ewald C. Palliative considerations in the
22
surgical treatment of spinal metastases: evaluation of posterolateral decompression 15 Page 15 of 22
1
combined with posterior instrumentation. J Cancer Res Clin Oncol 2012;138:301–10.
2
doi:10.1007/s00432-011-1100-3.
3
[14] Cho D-C, Sung J-K. Palliative surgery for metastatic thoracic and lumbar tumors using
4
posterolateral transpedicular approach with posterior instrumentation. Surg Neurol
5
2009;71:424–33. doi:10.1016/j.surneu.2008.02.049.
6
[15] Wise JJ, Fischgrund JS, Herkowitz HN, Montgomery D, Kurz LT. Complication, survival
7
rates, and risk factors of surgery for metastatic disease of the spine. Spine (Phila Pa 1976)
8
1999;24:1943–51.
9 10 11
[16] Bauer HC. Posterior decompression and stabilization for spinal metastases. Analysis of sixty-seven consecutive patients. J Bone Joint Surg Am 1997;79:514–22. [17] Pascal-Moussellard H, Broc G, Pointillart V, Siméon F, Vital JM, Sénégas J.
12
Complications of vertebral metastasis surgery. Eur Spine J 1998;7:438–44.
13
[18] Bridwell KH, Jenny AB, Saul T, Rich KM, Grubb RL. Posterior segmental spinal
14
instrumentation (PSSI) with posterolateral decompression and debulking for metastatic
15
thoracic and lumbar spine disease. Limitations of the technique. Spine (Phila Pa 1976)
16
1988;13:1383–94.
17
[19] Wang JC, Boland P, Mitra N, Yamada Y, Lis E, Stubblefield M, et al. Single-stage
18
posterolateral transpedicular approach for resection of epidural metastatic spine tumors
19
involving the vertebral body with circumferential reconstruction: results in 140 patients.
20
Invited submission from the Joint Section Meeting on Disorders. J Neurosurg Spine
21
2004;1:287–98. doi:10.3171/spi.2004.1.3.0287.
16 Page 16 of 22
1
[20] Wang JC, Boland P, Mitra N, Yamada Y, Lis E, Stubblefield M, et al. Single-stage
2
posterolateral transpedicular approach for resection of epidural metastatic spine tumors
3
involving the vertebral body with circumferential reconstruction: results in 140 patients.
4
Invited submission from the Joint Section Meeting on Disorders. J Neurosurg Spine
5
2004;1:287–98. doi:10.3171/spi.2004.1.3.0287.
6 7 8 9 10 11 12
[21] McPhee IB, Williams RP, Swanson CE. Factors influencing wound healing after surgery for metastatic disease of the spine. Spine (Phila Pa 1976) 1998;23:726-32-3. [22] Ghogawala Z, Mansfield FL, Borges LF. Spinal radiation before surgical decompression adversely affects outcomes of surgery for symptomatic metastatic spinal cord compression. Spine (Phila Pa 1976) 2001;26:818–24. [23] Sundaresan N, Rothman A, Manhart K, Kelliher K. Surgery for solitary metastases of the spine: rationale and results of treatment. Spine (Phila Pa 1976) 2002;27:1802–6.
13
[24] Yokogawa N, Murakami H, Demura S, Kato S, Yoshioka K, Hayashi H, et al.
14
Perioperative complications of total en bloc spondylectomy: adverse effects of
15
preoperative irradiation. PLoS One 2014;9:e98797. doi:10.1371/journal.pone.0098797.
16
[25] Blood AG, Sandoval MF, Burger E, Halverson-Carpenter K. Risk and Protective Factors
17
Associated with Surgical Infections among Spine Patients. Surg Infect (Larchmt)
18
2017;18:234–49. doi:10.1089/sur.2016.183.
19
[26] Sebaaly A, El Rachkidi R, Yaacoub JJ, Saliba E, Ghanem I. Management of spinal
20
infections in children with cerebral palsy. Orthop Traumatol Surg Res 2016;102:801–5.
21
doi:10.1016/j.otsr.2016.04.015.
17 Page 17 of 22
1
[27] Xing D, Ma J-X, Ma X-L, Song D-H, Wang J, Chen Y, et al. A methodological,
2
systematic review of evidence-based independent risk factors for surgical site infections
3
after spinal surgery. Eur Spine J 2013;22:605–15. doi:10.1007/s00586-012-2514-6.
4
[28] Graf K, Ott E, Vonberg R-P, Kuehn C, Schilling T, Haverich A, et al. Surgical site
5
infections--economic consequences for the health care system. Langenbeck’s Arch Surg
6
2011;396:453–9. doi:10.1007/s00423-011-0772-0.
7
[29] Kumar S, van Popta D, Rodrigues-Pinto R, Stephenson J, Mohammad S, Siddique I, et al.
8
Risk factors for wound infection in surgery for spinal metastasis. Eur Spine J
9
2015;24:528–32. doi:10.1007/s00586-013-3127-4.
10
[30] CDC. National Healthcare Safety Network: surgical site infection (SSI) event. 2014-01-01
11
[2014-03-25] Http//www, Cdc Gov/nhsn/pdfs/psc-manual/9pscssicurrent, Pdf 2010.
12
[31] Keam J, Bilsky MH, Laufer I, Shi W, Zhang Z, Tam M, et al. No association between
13
excessive wound complications and preoperative high-dose, hypofractionated, image-
14
guided radiation therapy for spine metastasis. J Neurosurg Spine 2014;20:411–20.
15
doi:10.3171/2013.12.SPINE12811.
16
[32] DiPaola CP, Saravanja DD, Boriani L, Zhang H, Boyd MC, Kwon BK, et al.
17
Postoperative Infection Treatment Score for the Spine (PITSS): construction and
18
validation of a predictive model to define need for single versus multiple irrigation and
19
debridement for spinal surgical site infection. Spine J 2012;12:218–30.
20
doi:10.1016/j.spinee.2012.02.004.
21 22
[33] Jansson K-A, Bauer HCF. Survival, complications and outcome in 282 patients operated for neurological deficit due to thoracic or lumbar spinal metastases. Eur Spine J 18 Page 18 of 22
1 2
2006;15:196–202. doi:10.1007/s00586-004-0870-6. [34] Atkinson RA, Davies B, Jones A, van Popta D, Ousey K, Stephenson J. Survival of
3
patients undergoing surgery for metastatic spinal tumours and the impact of surgical site
4
infection. J Hosp Infect 2016;94:80–5. doi:10.1016/j.jhin.2016.06.009.
5 6
[35] Awad SS. Adherence to surgical care improvement project measures and post-operative surgical site infections. Surg Infect (Larchmt) 2012;13:234–7. doi:10.1089/sur.2012.131.
7
19 Page 19 of 22
1
Figure Legends
2
Figure 1: Trends of decreasing of SSI incidence with publication date
3 4 5
Tables:
6
a) Table 1: Descriptive statistics N
Mean
Standard deviation
Age
297
60.97
10.9
Sex
297 53.9% Males; 46.1% Females
Smoking
292
29.5 % (86)
4.6%
Diabetes
276
14.1% (39)
3.4%
BMI
269
26.5
7.2
Known primary
296
69.6% (206)
ASA score
297
2.8
0.54
20 Page 20 of 22
Tokuhashi score
276
9.3
2.6
Number of involved vertebra
296
1.6
1
Preserved sphincteric function
289
82% (237)
History of Radiation therapy
295
34.9% (103)
Preoperative Hemoglobin levels 297
12.1
1.8
Off hours surgery
292
29.5% (86)
Surgical duration
297
208
115
Fused Levels
297
5.6
2.8
Operative bleeding (ml)
297
1175
1300
Infection
297
5.1% (15)
Dural Tear
297
8.4% (25)
Cardiovascular
297
4.4% (13)
Delirium
297
5.1% (15)
Pulmonary embolism
297
1.3 % (4)
Pneumonia
297
4.4% (13)
Other
297
11.1% (33)
Postoperative radiation therapy
263
69
Survival (days)
271
196
Complications:
1
235
* the numbers between parentheses represents the actual number of patients.
21 Page 21 of 22
1 2
b) Table 2: Evolution of SSI incidence according to publication date Author
Data collection Total Patients SSI SSI incidence
Bridwell [18]
1984
1987
25
3
12.0%
Pascal-Moussellard [17]
1982
1991
145
20
13.8%
McPhee [21]
1984
1998
75
15
20.0%
Sundaresan [23]
1986
1997
80
11
13.8%
Bauer [16]
1990
1994
67
11
16.4%
Wise [15]
1993
1996
80
8
10.0%
Janson [32]
1990
2001
282
35
12.4%
Demura [9]
1993
2003
113
8
7.1%
Wang [20]
1998
2002
140
16
11.4%
Cho [14]
2001
2006
21
2
9.5%
Walter [13]
2000
2010
57
2
3.5%
Bollen [12]
2001
2010
106
9
8.5%
Lee [5]
2005
2011
200
9
4.5%
de Ruiter [11]
2005
2011
60
6
10.0%
Quraishi [35]
2004
2012
289
13
4.5%
Park [10]
2006
2011
29
3
10.3%
Dea [36]
2009
2012
101
6
5.9%
Atkinson [33]
2009
2012
152
17
11.2%
Yokogawa [24]
2010
2013
50
4
8.0%
This study
2010
2016
374
19
5.1%
3 4 5
22 Page 22 of 22
S1529-9430(18)30004-4 https://doi.org/10.1016/j.spinee.2018.01.002 SPINEE 57566
To appear in:
The Spine Journal
Received date: Revised date: Accepted date:
19-9-2017 12-12-2017 10-1-2018
Please cite this article as: Amer Sebaaly, Daniel Shedid, Ghassan Boubez, Fahed Zairi, Michelle Kanhonou, Sung-Joo Yuh, Zhi Wang, Surgical site infection in spinal metastasis: incidence and risk factors, The Spine Journal (2018), https://doi.org/10.1016/j.spinee.2018.01.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.
1
Surgical site infection in spinal metastasis: incidence
2
and risk factors
3 4
Authors:
5
Amer SEBAALY (1, 2), Daniel SHEDID (3), Ghassan BOUBEZ (1), Fahed ZAIRI (3, 4),
6
Michelle KANHONOU (5), Sung-Joo YUH (3), Zhi WANG (1)
7 8 9 10
Affiliations: 1. Department of Orthopedic surgery, Spine Unit, Centre Hopitalier de l’Université de Montréal (CHUM), Montréal, Canada
11
2. Saint Joseph University, Faculty of medicine, Beirut, Lebanon
12
3. Department of Neurosurgery, Spine Unit, Centre Hopitalier de l’Université de Montréal
13
(CHUM), Montréal, Canada
14
4. Department of neurosurgery, Lille university hospital, Lille France
15
5. Research center of Montreal University, CR-CHUM, Montreal, Canada
16
Corresponding author:
17
Amer SEBAALY
18 19
Department of Orthopedic surgery, Spine Unit, Centre Hopitalier de l’Université de Montréal (CHUM), Montréal, Canada
20
Saint Joseph University, Faculty of medicine, Beirut, LEBANON
21
Hôpital Notre Dame, Rue Sherbrooke, Montréal, QC, H2L 4M1, Canada
22
Email: [email protected]
Tel +961 3 679 128 1 Page 1 of 22
1
Abstract:
2
Background: Surgical site infection (SSI) inspinal metastasis surgery represents the most
3
common postoperative surgical complication with high morbidity and mortality.
4
Objective: Evaluate the incidence of SSI in spinal metastasis surgery and its risk factors.
5
Study Design: Retrospective analysis of a prospective collected data.
6
Methods: Preoperative, operative and post-operative data were collected together with the
7
Modified Tokuhashi score and Frankel score at all time checkpoints. SSI was divided into
8
superficial and deep SSI as well as early (<90 days) and late SSI. Multiple logistic regression
9
analysis was performed to identify independent risk factors with p<0.05 as significance
10
threshold.
11
Results: 297 patients were included with an incidence of SSI of 5.1% (Superficial SSI:
12
3.4%; deep SSI: 1.7 %). Cervicothoracic surgery was associated with the highest incidence of
13
SSI while cervical surgery had the lowest incidence. Smoking, higher number of spinal
14
metastasis, elevated BMI and higher ASA score were the preoperative factors associated with
15
increased risk of SSI. Increased intraoperative blood loss and increased number of fixed vertebra
16
increased the SSI incidence. SSI increased hospital stay by a mean of 12 days. When all these
17
variables are analyzed in a multiple regression model, only surgical time ≥ 4 hours and ASA ≥ 3
18
were found to be independent risk factor for the occurrence of SSI.
19
Conclusion: This paper represents the largest series of spinal metastasis with a mean
20
incidence of SSI of 5.1%. Smoking, higher BMI, higher number of spinal metastasis, higher
21
ASA score, higher number of fused vertebra, intraoperative bleeding ≥ 2000 ml and neurological
22
deterioration are risk factors for SSI occurrence. Only ASA ≥ 3 and operative duration ≥ 4 hours 2 Page 2 of 22
1
are independent risk factors for this complication occurrence. Finally, SSI occurrence is
2
associated with increased hospital stay, increased 30 days mortality rate and decreased survival
3
rates.
4 5
Level of evidence: Level IV
6 7 8
Keywords: Spinal metastasis; Surgery; fixation; surgical site infection; acute infection; debridement.
3 Page 3 of 22
1
Introduction:
2
The axial skeleton is the third most frequent location of metastasis (after the liver and the
3
lungs) and the most common site of bone metastasis [1,2]. The incidence of spinal metastasis is
4
increasing with the improvement of treatment of primary tumors as well as the improvement of
5
palliative therapy [3]. Postmortem studies have shown that the incidence of spinal metastasis
6
ranges between 30-90% depending on the primary cancer type [4]. Ideal management of spinal
7
metastasis requires a multidisciplinary approach that involves radiation therapy (conventional /
8
Stereotactic Body Radiation Therapy (SBRT)), chemotherapy, immunotherapy (targeted
9
therapy), cement augmentation (vertebroplasty, kyphoplasty) and surgery (en bloc resection,
10
debulking or palliative surgery) [3,5]. The later is indicated in case of rapidly progressive
11
neurological deficit from neurological compression of the spinal cord by the tumor [6,7], the
12
need to establish diagnosis, radioresistant tumors, neurological deterioration under radiotherapy
13
or in the case of impending instability based in the Spinal Instability Neoplastic Score (SINS)
14
[2,8].
15
Even though recent reports have showed excellent functional outcomes [9], perioperative
16
complications of this spinal surgery are relatively high and range from 5.3 to 76.2% with an
17
average of 26.9% [3]. Surgical site infection (SSI) represents the most common postoperative
18
surgical complication [3,9]. The exact incidence of spinal SSI following metastasis surgery is
19
heterogenous in the literature and ranges between 3.51% and 20% with a mean of 10.22% [5,10–
20
25]. This incidence is eight to ten times higher than the mean SSI rate of 2.2% of spinal surgery
21
in general [1,9,26]. Even more, some authors found that spinal oncologic surgery is an
22
independent factor for SSI (OR=6.2, CI=1.7-22.3) [27]. Identically, the risk of mortality
23
increases by two folds with patients with SSI and by 1.5 fold for a stay in intensive care unit 4 Page 4 of 22
1
[28]. Several risk factors have been proposed for SSI: advanced age, low albumin levels,
2
multiple levels of surgery, use of preoperative steroids, closure with non resorbable skin suture,
3
surgery duration ≥ 4 hours, bleeding ≥ 3000 ml, preoperative neurological compromise, number
4
of instrumented vertebras, diabetes and preoperative radiation therapy [3,9,21,29]. Nonetheless,
5
these conclusions emerge from old retrospective studies with inhomogeneous inclusion criteria
6
and small sample sizes. Even more, these studies were not powered enough to conduct multiple
7
regression analysis to detect independent risk factor for SSI occurrence.
8
At our institution, a tertiary spine referral center, spinal metastasis surgery is one of the
9
most undergone surgeries (between 40-60 cases a year). Thus, since January 2010, a prospective
10
database was created to include all cases of spinal tumor surgery (primary and metastasis). The
11
objectives of this study are to analyze the incidence of SSI in this large database while trying to
12
analyze possible dependent and independent risk factors SSI as well as the effect of SSI on
13
hospital stay and overall mortality.
14 15
Material and methods:
16
This is a retrospective study of a prospective collected data from the spinal metastasis
17
surgery monocentric database. This study was approved by the institutional ethical board.
18
Inclusion criteria included age greater than 18 years old and patients operated for a spinal
19
metastasis of a known or unknown primary tumor. The main indication for operation was
20
epidural compression form the tumor and/or instability. Instability was assessed by the presence
21
of a burst fracture or severe osteolysis involving the anterior and posterior elements based on the
22
SINS score [8]. The data collection started on September 2011 (start of using the SINS criteria)
5 Page 5 of 22
1
till the end of December 2016. Patients were included consecutively and were operated by one of
2
the three senior surgeons (GB, DS, and ZW). Exclusion criteria were primary tumors (bone and
3
neurologic tumors), nonmalignant causes of spinal compression (benign tumors), as well as
4
patients whose neurological deficit was explained by another process (previous stroke). Patients
5
received postoperative adjuvant radiation therapy if deemed necessary by the radiation oncology
6
department. All patients received our prophylactic antibiotic protocol with cephazolin 1 to 2 g IV
7
30 min before incision repeated at 4 hours for lengthy procedures and cephazolin 1-2 g every 8
8
hours for 24h postoperatively (3 doses). If the patient has an allergy to penicillin, Vancomycin
9
replaces cephazolin. All patients received minimal 1L of irrigation before closure for these
10
complex procedures. Preoperative radiation therapy did not have a standardized protocol.
11
Patients had surgery within six months of the radiation therapy. As for postoperative radiation
12
therapy, the majority of the patients had a standard protocol depending on the primary tumor. All
13
patients received SBRT radiation with varying dose depending on the primary between 2 and 4
14
weeks postoperatively.
15
Data collection encompassed demographic data that included age, sex, and smoking
16
status. Tumor type, localization of the tumor as well as known primary status were also noted.
17
Preoperative data included Frankel score, modified Tokuhashi score, site(s) of compression,
18
number of spinal metastasis, history of radiation therapy, urinary sphincter function, preoperative
19
embolization of the spinal metastasis and preoperative hemoglobin level. Operative data included
20
type of procedure, bleeding volumes, ASA (American Society of Anesthesiologist) score, off-
21
hours surgery status (surgery starting between 17:00 and 8:00 the following day) as well as
22
surgery duration. Postoperative data is collected in the immediate postoperative period and at the
23
final follow-up and included postoperative complications (SSI, dural tear, pneumonia, delirium,
6 Page 6 of 22
1
cardiovascular complications including stroke, pulmonary embolism, others…), duration of stay,
2
last follow-up (FU) Frankel score, postoperative radiation and dose of the radiation therapy.
3
Survival analysis included survival rate and survival duration after surgery. This was last updated
4
on the start of the preparation of this manuscript (minimum 9 months follow-up for the last case).
5
Superficial SSI and deep SSI were defined according the CDC criteria [30]. Superficial
6
SSI is an infection involving only the dermis and the subcutaneous tissue without invading the
7
paraspinous fascia. Deep SSI is invades the fascia and is in direct contact with the implants and
8
is therefore more severe. When the patients presented with fever, erythema and tenderness, or
9
elevated CRP, high index of suspicion of deep SSI warranted MRI examination. In the absence
10
of clinical and imaging signs of invasion of the muscle fascia, the diagnosis of superficial SSI is
11
retained. Acute SSI was defined as SSI occurring during the first 90 days postoperatively while
12
late infection was defined as the one occurring after the 90 days of surgery. Patients were then
13
divided into two groups according to SSI incidence: patients in Group 1 had SSI, while patients
14
in group 2 didn’t (could have another complication). Moreover, some variables were
15
retrospectively collected for group 1 patients: number of Irrigation & Débridement (I&D),
16
antibiotherapy (duration and type of administration), removal of the instrumentation, and use of
17
vacuum assisted closure (VAC).
18
Statistical analysis is done using SPSS for Windows, Version 18.0. p values below 0.05
19
were considered statistically significant. Student t test was used to compare means. χ2 test was
20
used to compare binary variables and to calculate odds ratio. MANOVA test was used to identify
21
possible risk factors for SSI occurrence and finally a multiple logistic regression analysis was
22
made to identify independent risk factors.
23 7 Page 7 of 22
1
Results:
2
The database of primary and secondary tumors consisted of 458 cases and 297 patients
3
were included in this study since they were operated on for spinal metastasis between September
4
2011 and December 2016. Missing data represented less than 10% of the entire data.
5
Complication data was complete for all the cases. Incomplete data was considered missing
6
completely at random and thus ignored in statistical analysis. In this cohort, surgical indication
7
was neurologic compromise and/or instability according to the SINS criteria. The majority of
8
patients (90.5%) had posterior approach with decompression and fusion and 67% had
9
corpectomy. The remaining 9.5 % had anterior first corpectomy followed by posterior approach
10
for fixation. No case needed a flap for primary coverage. Demographic, preoperative, operative
11
and postoperative data are summarized in table 1. Around 8.7% of patients had neurologic
12
deterioration following surgery whereas 68.2% had no change in their Frankel score and 23.2%
13
had an improvement of minimum one Frankel score.
14
Mean SSI rate was 5.1% (15 patients). Superficial SSI had and incidence of 3.4% (10
15
patients) while deep SSI had an incidence of 1.7 % (number of cases = 5). Patients with deep SSI
16
were debrided by a mean of 1.2 debridements one patient requiring two debridements). All but
17
one patient presented with acute or subacute infection (mean time to infection 26.7 days) with
18
only one patient presenting with infection 8 months post-operatively. Staphylococcus aureus was
19
incriminated in the majority of cases (73%). One infection was caused by more than one
20
organism (Echerichia coli). Vancomycin was the most used IV antibiotic with a mean duration
21
of IV antibiotic was 16.2 days. One patient (delayed infection case) required instrumentation
22
removal and VAC closure, followed by instrumentation afterwards and flap coverage for closure.
23
There was a tendency for a seasonal effect with the lowest infection rate (2.6%) recorded in 8 Page 8 of 22
1
summer (July-September) whereas highest incidence rate was recorded in winter (January-
2
March; 8%; p=0.13). Cervicothoracic surgery was associated with the highest incidence of SSI
3
(12.5%) while cervical surgery had the lowest incidence (2.22%), with a tendency to statistical
4
significance when compared to dorsal surgery (SSI incidence= 5.92%; p=0.15). Lumbar surgery
5
has an SSI incidence of 2.94%.
6
When comparing patients with SSI with patients with no SSI (group 1 vs group 2), higher
7
smoking incidence was seen in group 1 (57.2% vs 23%; p=0.04). Mean number of involved
8
vertebrae was increased in group 1 (2.4) compared to group 2 (1.6 (p=0.03)). There was no
9
difference in mean Tokushashi score and no effect of preoperative diabetes status. Higher BMI
10
was noted in the group 1 (29.3 vs 24.5; p=0.07). Fifty seven percent of patients with SSI had a
11
history of radiation therapy compared to only 33% of patients with no SSI (p=0.009). Likewise,
12
ASA score was higher in group 1 (3) compared to group 2 (2.76; p=0.04) as well as for operative
13
bleeding (mean group 1= 2126 ml; mean group 2: 1224 ml; p=0.005). The SSI group had a
14
higher number of fixed vertebra (7.8 vs 5.5; p=0.03). SSI increased hospital stay by a mean of 12
15
days (25 in group 1 compared to 13 in group 2; p=0.002). Postoperative Ambulatory status did
16
not affect the occurrence of SSI (Non ambulatory: SSI rate of 6.1%; ambulatory: 3.92%; p =
17
0.4). Finally, mean survival with patients in group 1 was 120 days compared to 202 days in
18
group 2 (p=0.02).
19
Odds ratios were then calculated for the dependent variables. Smoking increased the risk
20
of SSI by 2.4 (1.06-7.2; p=0.04), ASA score ≥ 3 increased the odds of SSI by 1.1 (1.03-1.2;
21
p=0.02) while operative bleeding ≥ 2000 ml increased SSI by 2.66 (1.1-7; p=0.04). Neurological
22
deterioration was associated with an increased risk of SSI (OR= 4.2 (1.4-12.8; p=0.02). On the
23
other hand, the occurrence of any other complication (aside from SSI) did not increase the risk of 9 Page 9 of 22
1
SSI (OR=1.5 (0.5-3.9; p=0.5). Finally, surgical duration more than 4 hours was associated with
2
an increased risk of SSI (OR=3.76 (1.2-11.6; p=0.015). When all these variables are analyzed in
3
a multiple regression model, only duration ≥ 4 hours and ASA ≥ 3 were independent risk factor
4
for the occurrence of SSI.
5 6
Discussion:
7
SSI is a serious complication following spinal surgery in general and metastasis surgery
8
in particular. This study reports the incidence of SSI in a homogenous monocentric population of
9
patients with spinal metastasis surgery. This cohort represents the largest cohort in the literature
10
evaluating this complication with nearly 1.5 times the number of the largest reported series [5]).
11
The incidence of SSI was 5.1%, approximately half of the reported mean of 10.22% in the
12
literature [3]. Moreover, this is the first paper to study the effect of primary tumor type and
13
location on the occurrence of SSI with no effect of tumor type and a tendency of higher SSI in
14
dorsal spine compared to cervical spine [3]. Smoking, higher number of spinal metastasis, higher
15
ASA score, intraoperative bleeding ≥ 2000 ml and neurological deterioration were all found to
16
be risk factors for SSI occurrence. Only ASA ≥ 3 and operative duration ≥ 4 hours are
17
independent risk factors for this complication occurrence. Finally, SSI occurrence is associated
18
with increased hospital stay, increased 30 days mortality rate and decreased survival rates.
19
The incidence of SSI in this study is one of the lowest in the modern spine literature with
20
a 5.1% incidence of SSI (1.7% for deep SSI and 3.4% for superficial SSI). Atkinson et al found a
21
similar rate of deep (2%) with a higher superficial SSI (9.2%) with superficial to deep ratio of
22
4.6/1 [1]. In a similar manner, Demura et al found higher deep SSI incidence of 7.1% [9] and
10 Page 10 of 22
1
other studies found higher rate of deep SSI requiring reoperation (9.45%) [2]. This decreasing
2
rate of SSI along the years could be attributed to better infection prevention protocols and better
3
postoperative rehabilitations (Table 2; Figure 1). Reoperation rate for SSI was 2%. This finding
4
is much less compared to other studies in the literature (4.4 % reoperation rate in the study of
5
Quraishi et al [2]). Even more, I&D is needed in about one half of all SSI cases and all cases of
6
deep SSI, findings similar to the available literature [1,9]. This increased need for reoperation
7
exposes this particular population of frail patients to risk of recurrent anesthesia as well as other
8
surgical complications.
9
Many studies have evaluated the possible risk factors for the development of SSI.
10
Smoking is a known risk factor for SSI in general surgery and general spinal surgery but no
11
study has been able to show its correlation to SSI in spinal metastasis [25,31]. This is the first
12
study to find that preoperative smoking is a risk factor for SSI (OR= 2.4). On the other hand,
13
patients’ comorbidities are known risk factor for spinal SSI and constitute a category in the
14
PITSS scoring system for the need of I&D in deep spinal SSI [32]. Higher ASA scores ≥ 3,
15
denoting the presence of more co-morbidities, was found to be one important risk factor for the
16
occurrence of this complication. Even more, a lengthier surgery with higher bleeding volumes
17
are known risk factors for SSI as shown in this study [21,33]. Albeit neurologic deficits are
18
associated with higher risk of SSI in deformity or trauma surgery [26], there was no association
19
between ambulatory status and SSI as found by other authors in the literature [9,29]. Only
20
neurological deterioration after surgery was associated with increased SSI rates (OR=4.2).
21
Finally, the most controversial risk factor for SSI is preoperative radiation therapy. Classically,
22
this neoadjuvant therapy has negative effects on wound repair due to decreased vascularity,
23
hypoxia and impairment of tissue repair [3,9]. Some authors related this complication to the dose
11 Page 11 of 22
1
of radiation with increased complications with doses greater than 40 Gy [24]. Others found that
2
preoperative radiotherapy was associated with higher risk of deep SSI with an OR ranging from
3
3 to 41 [9,22,23]. Nonetheless, no increase of wound complications and SSI with the use of
4
radiation therapy was found in the study of Keam et al [31]. Similarly, we have not found
5
preoperative radiation therapy to be a risk factor of SSI.
6
The consequences of SSI in spinal metastasis surgeries are also debated. Duration of
7
hospitalization with patients with SSI increases by 100% (12 days) and is around 4 weeks (25
8
days). This finding is similar to the findings of other authors with a mean hospitalization duration
9
of 32 days [1]. To add to that, SSI affect survival after spinal metastasis surgery. We found a
10
decrease of survival by 4 months; same results were found by other authors with a one year
11
survivorship of 42% [34]. The study of Atkinson et al. didn’t have the statistical power to
12
confirm the impact of SSI on mortality. This decrease of survival could be associated with the
13
occurrence of SSI per se or the increased frailty of the patients who develop SSI: higher ASA
14
and lower Modified Tokuhashi scores. In a recent study, authors found that 38 to 75% of deaths
15
in patients with SSI are attributable to SSI itself [35]. Due to the small number of cases, the true
16
effect of deep versus superficial SSI on mortality could not be studied in this series and probably
17
needs multicentric efforts to increase the population number.
18
This study has several limitations. First of all, this is a prospectively collected database
19
that could have some missing data. Data collection, even thorough designed to be exhaustive,
20
could be imperfect. We considered missing data to be totally at random. Second of all, the low
21
rate of SSI could prevent us from detecting correlations and statistical difference on some
22
variables. Multicentric data pooling could help to increase the included number of patients and
12 Page 12 of 22
1
overcome this limitation. Nonetheless, this study represents the largest monocentric series in the
2
modern literature and drawn conclusions have solid clinical and statistical power.
3 4
Conclusion:
5
The incidence of SSI in the largest population of patients with spinal metastasis surgery
6
reaches 5.1%. Primary tumor type and location have no effect on SSI incidence. Smoking, higher
7
BMI, higher number of spinal metastasis, higher ASA score, higher number of fused vertebra,
8
intraoperative bleeding ≥ 2000 ml and neurological deterioration are risk factors for SSI
9
occurrence. Only ASA ≥ 3 and operative duration ≥ 4 hours are independent risk factors for this
10
complication occurrence. Finally, SSI occurrence is associated with increased hospital stay,
11
increased 30 days mortality rate and decreased survival rates.
13 Page 13 of 22
1 2
References: [1]
Atkinson RA, Jones A, Ousey K, Stephenson J. Management and cost of surgical site
3
infection in patients undergoing surgery for spinal metastasis. J Hosp Infect 2017;95:148–
4
53. doi:10.1016/j.jhin.2016.11.016.
5
[2]
Quraishi N a., Rajabian A, Spencer A, Arealis G, Mehdian H, Boszczyk BM, et al.
6
Reoperation rates in the surgical treatment of spinal metastases. Spine J 2015;15:S37–43.
7
doi:10.1016/j.spinee.2015.01.005.
8
[3]
9
DB. Perioperative Complications of Spinal Metastases Surgery. Clin Spine Surg
10 11
2017;30:4–13. doi:10.1097/BSD.0000000000000484. [4]
12 13
Luksanapruksa P, Buchowski JM, Zebala LP, Kepler CK, Singhatanadgige W, Bumpass
Wong DA, Fornasier VL, MacNab I. Spinal metastases: the obvious, the occult, and the impostors. Spine (Phila Pa 1976) 1990;15:1–4.
[5]
Lee BH, Park J-O, Kim H-S, Park Y-C, Lee H-M, Moon S-H. Perioperative complication
14
and surgical outcome in patients with spine metastases: retrospective 200-case series in a
15
single institute. Clin Neurol Neurosurg 2014;122:80–6.
16
doi:10.1016/j.clineuro.2014.04.025.
17
[6]
Dodwad S-NM, Savage J, Scharschmidt TJ, Patel A. Evaluation and treatment of spinal
18
metastatic disease. Cancer Treat Res 2014;162:131–50. doi:10.1007/978-3-319-07323-
19
1_7.
20 21
[7]
Patchell RA, Tibbs PA, Regine WF, Payne R, Saris S, Kryscio RJ, et al. Direct decompressive surgical resection in the treatment of spinal cord compression caused by
14 Page 14 of 22
1
metastatic cancer: a randomised trial. Lancet (London, England) 2005;366:643–8.
2
doi:10.1016/S0140-6736(05)66954-1.
3
[8]
Fisher CG, DiPaola CP, Ryken TC, Bilsky MH, Shaffrey CI, Berven SH, et al. A novel
4
classification system for spinal instability in neoplastic disease: an evidence-based
5
approach and expert consensus from the Spine Oncology Study Group. Spine (Phila Pa
6
1976) 2010;35:E1221-9. doi:10.1097/BRS.0b013e3181e16ae2.
7
[9]
Demura S, Kawahara N, Murakami H, Nambu K, Kato S, Yoshioka K, et al. Surgical Site
8
Infection in Spinal Metastasis. Spine (Phila Pa 1976) 2009;34:635–9.
9
doi:10.1097/BRS.0b013e31819712ca.
10
[10] Park S-J, Lee C-S, Chung S-S. Surgical results of metastatic spinal cord compression
11
(MSCC) from non-small cell lung cancer (NSCLC): analysis of functional outcome,
12
survival time, and complication. Spine J 2016;16:322–8.
13
doi:10.1016/j.spinee.2015.11.005.
14
[11] de Ruiter GCW, Lobatto DJ, Wolfs JF, Peul WC, Arts MP. Reconstruction with
15
expandable cages after single- and multilevel corpectomies for spinal metastases: a
16
prospective case series of 60 patients. Spine J 2014;14:2085–93.
17
doi:10.1016/j.spinee.2013.12.029.
18
[12] Bollen L, de Ruiter GCW, Pondaag W, Arts MP, Fiocco M, Hazen TJT, et al. Risk factors
19
for survival of 106 surgically treated patients with symptomatic spinal epidural
20
metastases. Eur Spine J 2013;22:1408–16. doi:10.1007/s00586-013-2726-4.
21
[13] Walter J, Reichart R, Waschke A, Kalff R, Ewald C. Palliative considerations in the
22
surgical treatment of spinal metastases: evaluation of posterolateral decompression 15 Page 15 of 22
1
combined with posterior instrumentation. J Cancer Res Clin Oncol 2012;138:301–10.
2
doi:10.1007/s00432-011-1100-3.
3
[14] Cho D-C, Sung J-K. Palliative surgery for metastatic thoracic and lumbar tumors using
4
posterolateral transpedicular approach with posterior instrumentation. Surg Neurol
5
2009;71:424–33. doi:10.1016/j.surneu.2008.02.049.
6
[15] Wise JJ, Fischgrund JS, Herkowitz HN, Montgomery D, Kurz LT. Complication, survival
7
rates, and risk factors of surgery for metastatic disease of the spine. Spine (Phila Pa 1976)
8
1999;24:1943–51.
9 10 11
[16] Bauer HC. Posterior decompression and stabilization for spinal metastases. Analysis of sixty-seven consecutive patients. J Bone Joint Surg Am 1997;79:514–22. [17] Pascal-Moussellard H, Broc G, Pointillart V, Siméon F, Vital JM, Sénégas J.
12
Complications of vertebral metastasis surgery. Eur Spine J 1998;7:438–44.
13
[18] Bridwell KH, Jenny AB, Saul T, Rich KM, Grubb RL. Posterior segmental spinal
14
instrumentation (PSSI) with posterolateral decompression and debulking for metastatic
15
thoracic and lumbar spine disease. Limitations of the technique. Spine (Phila Pa 1976)
16
1988;13:1383–94.
17
[19] Wang JC, Boland P, Mitra N, Yamada Y, Lis E, Stubblefield M, et al. Single-stage
18
posterolateral transpedicular approach for resection of epidural metastatic spine tumors
19
involving the vertebral body with circumferential reconstruction: results in 140 patients.
20
Invited submission from the Joint Section Meeting on Disorders. J Neurosurg Spine
21
2004;1:287–98. doi:10.3171/spi.2004.1.3.0287.
16 Page 16 of 22
1
[20] Wang JC, Boland P, Mitra N, Yamada Y, Lis E, Stubblefield M, et al. Single-stage
2
posterolateral transpedicular approach for resection of epidural metastatic spine tumors
3
involving the vertebral body with circumferential reconstruction: results in 140 patients.
4
Invited submission from the Joint Section Meeting on Disorders. J Neurosurg Spine
5
2004;1:287–98. doi:10.3171/spi.2004.1.3.0287.
6 7 8 9 10 11 12
[21] McPhee IB, Williams RP, Swanson CE. Factors influencing wound healing after surgery for metastatic disease of the spine. Spine (Phila Pa 1976) 1998;23:726-32-3. [22] Ghogawala Z, Mansfield FL, Borges LF. Spinal radiation before surgical decompression adversely affects outcomes of surgery for symptomatic metastatic spinal cord compression. Spine (Phila Pa 1976) 2001;26:818–24. [23] Sundaresan N, Rothman A, Manhart K, Kelliher K. Surgery for solitary metastases of the spine: rationale and results of treatment. Spine (Phila Pa 1976) 2002;27:1802–6.
13
[24] Yokogawa N, Murakami H, Demura S, Kato S, Yoshioka K, Hayashi H, et al.
14
Perioperative complications of total en bloc spondylectomy: adverse effects of
15
preoperative irradiation. PLoS One 2014;9:e98797. doi:10.1371/journal.pone.0098797.
16
[25] Blood AG, Sandoval MF, Burger E, Halverson-Carpenter K. Risk and Protective Factors
17
Associated with Surgical Infections among Spine Patients. Surg Infect (Larchmt)
18
2017;18:234–49. doi:10.1089/sur.2016.183.
19
[26] Sebaaly A, El Rachkidi R, Yaacoub JJ, Saliba E, Ghanem I. Management of spinal
20
infections in children with cerebral palsy. Orthop Traumatol Surg Res 2016;102:801–5.
21
doi:10.1016/j.otsr.2016.04.015.
17 Page 17 of 22
1
[27] Xing D, Ma J-X, Ma X-L, Song D-H, Wang J, Chen Y, et al. A methodological,
2
systematic review of evidence-based independent risk factors for surgical site infections
3
after spinal surgery. Eur Spine J 2013;22:605–15. doi:10.1007/s00586-012-2514-6.
4
[28] Graf K, Ott E, Vonberg R-P, Kuehn C, Schilling T, Haverich A, et al. Surgical site
5
infections--economic consequences for the health care system. Langenbeck’s Arch Surg
6
2011;396:453–9. doi:10.1007/s00423-011-0772-0.
7
[29] Kumar S, van Popta D, Rodrigues-Pinto R, Stephenson J, Mohammad S, Siddique I, et al.
8
Risk factors for wound infection in surgery for spinal metastasis. Eur Spine J
9
2015;24:528–32. doi:10.1007/s00586-013-3127-4.
10
[30] CDC. National Healthcare Safety Network: surgical site infection (SSI) event. 2014-01-01
11
[2014-03-25] Http//www, Cdc Gov/nhsn/pdfs/psc-manual/9pscssicurrent, Pdf 2010.
12
[31] Keam J, Bilsky MH, Laufer I, Shi W, Zhang Z, Tam M, et al. No association between
13
excessive wound complications and preoperative high-dose, hypofractionated, image-
14
guided radiation therapy for spine metastasis. J Neurosurg Spine 2014;20:411–20.
15
doi:10.3171/2013.12.SPINE12811.
16
[32] DiPaola CP, Saravanja DD, Boriani L, Zhang H, Boyd MC, Kwon BK, et al.
17
Postoperative Infection Treatment Score for the Spine (PITSS): construction and
18
validation of a predictive model to define need for single versus multiple irrigation and
19
debridement for spinal surgical site infection. Spine J 2012;12:218–30.
20
doi:10.1016/j.spinee.2012.02.004.
21 22
[33] Jansson K-A, Bauer HCF. Survival, complications and outcome in 282 patients operated for neurological deficit due to thoracic or lumbar spinal metastases. Eur Spine J 18 Page 18 of 22
1 2
2006;15:196–202. doi:10.1007/s00586-004-0870-6. [34] Atkinson RA, Davies B, Jones A, van Popta D, Ousey K, Stephenson J. Survival of
3
patients undergoing surgery for metastatic spinal tumours and the impact of surgical site
4
infection. J Hosp Infect 2016;94:80–5. doi:10.1016/j.jhin.2016.06.009.
5 6
[35] Awad SS. Adherence to surgical care improvement project measures and post-operative surgical site infections. Surg Infect (Larchmt) 2012;13:234–7. doi:10.1089/sur.2012.131.
7
19 Page 19 of 22
1
Figure Legends
2
Figure 1: Trends of decreasing of SSI incidence with publication date
3 4 5
Tables:
6
a) Table 1: Descriptive statistics N
Mean
Standard deviation
Age
297
60.97
10.9
Sex
297 53.9% Males; 46.1% Females
Smoking
292
29.5 % (86)
4.6%
Diabetes
276
14.1% (39)
3.4%
BMI
269
26.5
7.2
Known primary
296
69.6% (206)
ASA score
297
2.8
0.54
20 Page 20 of 22
Tokuhashi score
276
9.3
2.6
Number of involved vertebra
296
1.6
1
Preserved sphincteric function
289
82% (237)
History of Radiation therapy
295
34.9% (103)
Preoperative Hemoglobin levels 297
12.1
1.8
Off hours surgery
292
29.5% (86)
Surgical duration
297
208
115
Fused Levels
297
5.6
2.8
Operative bleeding (ml)
297
1175
1300
Infection
297
5.1% (15)
Dural Tear
297
8.4% (25)
Cardiovascular
297
4.4% (13)
Delirium
297
5.1% (15)
Pulmonary embolism
297
1.3 % (4)
Pneumonia
297
4.4% (13)
Other
297
11.1% (33)
Postoperative radiation therapy
263
69
Survival (days)
271
196
Complications:
1
235
* the numbers between parentheses represents the actual number of patients.
21 Page 21 of 22
1 2
b) Table 2: Evolution of SSI incidence according to publication date Author
Data collection Total Patients SSI SSI incidence
Bridwell [18]
1984
1987
25
3
12.0%
Pascal-Moussellard [17]
1982
1991
145
20
13.8%
McPhee [21]
1984
1998
75
15
20.0%
Sundaresan [23]
1986
1997
80
11
13.8%
Bauer [16]
1990
1994
67
11
16.4%
Wise [15]
1993
1996
80
8
10.0%
Janson [32]
1990
2001
282
35
12.4%
Demura [9]
1993
2003
113
8
7.1%
Wang [20]
1998
2002
140
16
11.4%
Cho [14]
2001
2006
21
2
9.5%
Walter [13]
2000
2010
57
2
3.5%
Bollen [12]
2001
2010
106
9
8.5%
Lee [5]
2005
2011
200
9
4.5%
de Ruiter [11]
2005
2011
60
6
10.0%
Quraishi [35]
2004
2012
289
13
4.5%
Park [10]
2006
2011
29
3
10.3%
Dea [36]
2009
2012
101
6
5.9%
Atkinson [33]
2009
2012
152
17
11.2%
Yokogawa [24]
2010
2013
50
4
8.0%
This study
2010
2016
374
19
5.1%
3 4 5
22 Page 22 of 22