European Spine Journal (2023) 32:75–83

https://doi.org/10.1007/s00586-022-07339-z

ORIGINAL ARTICLE

Percutaneous pedicle screw fixation without arthrodesis of 368

thoracolumbar fractures: long‑term clinical and radiological outcomes
in a single institution
Sokol Trungu1,2 · Luca Ricciardi2 · Stefano Forcato1 · Amadeo Piazza2 · Giancarlo D’Andrea3 · Filippo Maria Polli4 ·
Marco Cimatti2 · Alessandro Frati5 · Massimo Miscusi2 · Antonino Raco2

Received: 4 April 2022 / Revised: 15 July 2022 / Accepted: 20 July 2022 / Published online: 3 August 2022
© The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2022

Abstract
Purpose Traumatic thoracolumbar (TL) fractures are the most common vertebral fractures. Although a consensus on the
preferred treatment is missing, percutaneous pedicle screw fixation (PPSF) has been progressively accepted as treatment
option, since it is related to lower soft tissues surgical-injury and perioperative complications rate. This study aims to evaluate
the long-term clinical–radiological outcomes after PPSF for TL fractures at a single tertiary academic hospital.
Methods This is a retrospective cohort study. Back pain was obtained at preoperative, postoperative and final follow-up
using Visual Analog Scale. Patient-reported outcomes, the Oswestry Disability Index and the 36-Item Short Form, were
obtained to asses disability during follow-up. Radiological measures included Cobb angle, mid-sagittal index, sagittal index
(SI) and vertebral body height loss. A multivariate regression analysis on preoperative radiological features was performed
to investigate independent risk factors for implant failure.
Results A total of 296 patients with 368 TL fractures met inclusion criteria. Mean follow-up was 124.3 months. The clini-
cal and radiological parameters significantly improved from preoperative to last follow-up measurements. The multivariate
analysis showed that Cobb angle (OR = 1.3, p < 0.001), SI (OR = 1.5, p < 0.001) and number of fractures (OR = 1.1, p = 0.05),
were independent risk factors for implant failure. The overall complication rate was 5.1%, while the reoperation rate for
implant failure was 3.4%.
Conclusions In our case series, PPSF for TL injuries demonstrated good long-term clinical-radiological outcomes, along
with low complication and reoperation rates. Accordingly, PPSF could be considered as a valuable treatment option for
neurologically intact patients with TL fractures. Additionally, in this cohort, number of fractures ≥ 2, Cobb angle ≥ 15° and
sagittal index ≥ 21° were independent risk factors for implant failure.

Keywords Thoracolumbar fracture · Percutaneous pedicle screw fixation · Burst fracture · Minimally invasive surgery ·
Posterior short-segment fixation

Abbreviations
TL Thoracolumbar
PPSF Percutaneous pedicle screw fixation
CT Computed tomography
* Sokol Trungu
[email protected] MIS Minimally invasive surgery
MSI Mid-sagittal index
1
Neurosurgery Unit, Cardinale G. Panico Hospital, Via MRI Magnetic resonance imaging
Fratelli Peluso 8, 73039 Tricase, Italy
ODI Oswestry Disability Index
2
N.E.S.M.O.S. Department, Sant’Andrea Hospital, Sapienza” PLC Posterior ligamentous complex
University of Rome, Rome, Italy SI Sagittal index
3
Neurosurgery Unit, F. Spaziani Hospital, Frosinone, Italy VAS Visual analog scale
4
Department of Neurosurgery, Fondazione Policlinico VBHL Vertebral body height loss
Universitario Agostino Gemelli - IRCCS, Rome, Italy
5
Neurosurgery, Neuromed IRCCS, Pozzilli, Italy

13
Vol.:(0123456789)
76 European Spine Journal (2023) 32:75–83

Introduction Patient population

Traumatic thoracolumbar (TL) fractures represent the Patients admitted at our tertiary academic referral center
most common fractures of the spine. Up to 50–60% of for traumatic TL fractures who underwent PPSF, between
these fractures involve the TL junction (T10-L2), 25–45% March 2007 and March 2015 (8 years), were considered for
the thoracic spine, and 10–14% the lumbosacral spine [1, eligibility.
2]. The compression-burst fractures (type A, AOSpine Inclusion criteria were: traumatic TL fractures classi-
Thoracolumbar Spine Injury Classification System [3, 4]) fied according to the AOSpine Thoracolumbar Spine Injury
are the most common (60–70%) subtype, followed by dis- Classification System; neurologically intact (ASIA E);
traction injuries (type B—14–21%), and rotational injuries follow-up > 6 years.
(type C—4–16%) [5]. Exclusion criteria were: incomplete clinical and radiolog-
Minimally invasive surgery (MIS) has been progres- ical data; history of severe osteoporosis (DEXA—T-score
sively accepted as a treatment option for spine fractures. of − 2.5 or lower and presence of one or more fragility frac-
Several studies have compared posterior percutaneous tures, as defined by the World Health Organization); previ-
screw fixation (PPSF) techniques with traditional open ous TL surgery; non-traumatic fractures; < 6 years follow-up.
surgery, concluding that PPSF is associated with shorter
surgical duration and hospital stay, lower intraoperative Surgical technique
blood loss and infection rates, and lower surgical-related
injury to spinal muscles [6, 11]. On the other hand, no A standard technique for percutaneous screw placement was
significative differences were reported concerning radio- performed in all patients. A convergent and straightforward
logical and clinical outcomes of these techniques [12, 13]. pedicle screw trajectory is preferred. After insertion of all
Recently, the Congress of Neurological Surgeons (CNS) screws, contouring of the rods is performed and these are
guidelines have recommended to carefully consider the inserted and locked in place.
role of arthrodesis in these procedures, since osteosynthe- Patients, with monosegmental burst fractures, were
sis alone seems to provide similar clinical and radiological treated with short segment fixation (6 screws) using an
outcomes [14]. Posterior short-segment fixation has dem- instrumentation system involving one vertebra above, one
onstrated as effective in restoring spinal stability, correct- below and at the fractured vertebra (in which either mon-
ing kyphotic deformity, and indirectly decompressing the olateral or bilateral pedicle screws were inserted according
spinal canal in TL fractures [15–20]. However, there are to local anatomy and pedicles status). Segmental lordosis
no conclusive long-term clinical-radiological data in TL restoration was always pursued in the case of segmental
fractures management. kyphosis with percutaneous distraction.
The aim of this study was to evaluate the long-term In cases of multiple fractures or type B fractures, the num-
clinical and radiological outcomes of 296 neurologically ber of levels involved with pedicle screws depended on the
intact patients with 368 TL fractures who underwent PPSF type, level, number of fractures and deformity grade. Can-
without arthrodesis at a single institution. To the best of nulated polyaxial screws were used in all patients. Screws’
our knowledge, our data are based on the largest cohort of sizes ranged in length and diameter, from 30 to 55 mm and
patients, from a single-institution, and the longest clinical from 4.5 to 7.5 mm, respectively. Titanium rods sizes ranged
and radiological follow-up. in diameter from 5 to 6.5 mm. Different percutaneous instru-
mentation systems were used over years. Bracing was never
prescribed after surgery in any case. Implant removal was
not perfomed in any patients of this study.
Methods
Clinical outcomes
Study design
General and neurological conditions, as well as the quality of
This is a retrospective cohort study from a single institu- life, were evaluated at admittance (preoperative parameters),
tion. According to the study design and national and insti- 6 weeks, one year, 6 years after surgery, and at final follow-
tutional guidelines, the ethical committee approval was not up visit, using a ten-points itemized visual analog scale
required. At the time of hospitalization, all the patients (VAS) for low-back pain. Patient reported outcomes were
provided their written informed consent for surgery and evaluated at 6 weeks, one year, 6 years after surgery and final
data management for scientific purposes. This study agrees follow-up visit to asses disability using the Oswestry Dis-
with the WMA Helsinki declaration of Human Rights. ability Index (ODI) score and the short-form SF-36 score.

13
European Spine Journal (2023) 32:75–83 77

Trauma types and mechanisms (motor vehicle col- Results

lision, sports trauma, work/domestic fall, diving) were
recorded. Thoracolumbar AOSpine Injury (TL AOSIS) Patient and operative characteristics
[21] classification was recorded for each fracture. Intra-,
post-operative and during follow-up complications were A total of 296 patients who underwent PPSF for 368 trau-
collected and classified as major and minor as described matic TL fractures during the study period met the inclu-
by Glassman et al. [22]. sion criteria and were included for data analysis.
There were 126 (42.6%) women and 170 (57.4%)
men. The mean age at the time of surgery was
Radiological outcomes 46.2 ± 19.3 years (range 15–86). The mean follow-up was
124.3 ± 26.1 months (range 78–174). The most common
Preoperative X-rays and computed tomography (CT) scan co-morbidity was cardiovascular diseases (49.3%), fol-
were retrieved from institutional picture archiving and lowed by diabetes mellitus (33.4%), obesity (29.4%) and
communication system (PACS). Preoperative magnetic respiratory diseases (13.8%). Ninety-six patients (32.4%)
resonance imaging (MRI) was performed only in cases were smokers. Ninety-two patients were classified as ASA
suspicious for posterior ligamentous complex injury. Class I (31.1%), 149 (50.3%) as Class II, 45 (15.2%) as
The following radiological parameters were calcu- Class III and 10 (3.4%) as Class IV.
lated and collected: the mid-sagittal index (MSI)—the The main traumatic mechanism was car/motorbike
percentage of loss of the anteroposterior diameter of the accident (66.2%), followed by professional-related inju-
spinal canal at the fractured vertebra level related to the ries (13.5%), sport activities (8.4%), domestic accidents
same mean value at the overlying and underlying levels (7.8%) and diving (4.1%).
[23]; the Cobb angle—the angle between a line drawn According to the recent AOSpine Classification,
parallel to the superior end plate of one vertebra above the cohort included 76 A3-type fractures (20.7%), 234
the fracture and a line drawn parallel to the inferior end A4-type fractures (63.6%), 36 B1 type (9.8%), 16 B2 type
plate of the vertebra one level below the fracture [24]; (4.3%) and 6 B3 type fractures (1.6%). The most frequent
the sagittal index (SI)—the measurement of segmental fractured level was the thoracolumbar junction T10-L2 in
kyphosis at the level of a mobile segment (1 vertebra 179 patients (53.3%), followed by thoracic spine T4–T9
and 1 disc) adjusted for the baseline sagittal contour at in 108 (32.1%), and lumbar spine L3–L5 in 49 (14.6%).
that level [25] and vertebral body height loss (VBHL)— Two hundred and thirty-three patients (78.7%) had one
measured by anterior/posterior body height compression fracture, whereas 54 (18.2%) and 9 patients (3.1%) had
ratio [24]. two and three fractures, respectively. The mean TL AOSIS
Screw misplacement was evaluated on postoperative was 5 (range 3–8).
CT scan and classified as by Gertzbein et al. [26]. The The mean length of surgery was 54.1 ± 23.8 min
presence of screw breakage, screw pullout, peri-implant (range 30–150), with an average of 46.6 ± 15.2 ml (range
loosening, and rod breakage were considered as criteria 30–110 ml) of estimated blood loss (EBL). The mean
for implant failure. length of stay (LOS) was 3 days (2.6 ± 3.6; range 2–40)
and the mean time of postoperative mobilization was
2 days (1.5 ± 3.1; range 2–30). No intraoperative compli-
Statistical analysis cations were recorded. Two hundred and fifty-five patients
(86.1%) were discharged to home, while 13.9% of patients
Values were reported as mean ± standard deviation. The to a rehabilitation unit.
t-Student test was used to compare the quantitative con- Patients demographic and operative characteristics are
tinuous variables. Fisher's exact test (2-sided) was used summarized in Tables 1 and 2.
instead to compare the categorical variables. Statistical
significance was pre-determined at an alpha of 0.05. Uni-
variate and Multivariate binary logistic regression analysis Clinical and radiological outcomes
with odds ratio (OR) was performed to evaluate independ-
ent radiological risk factors for implant failure. Further- Comparing mean preoperative and last follow-up val-
more, for the independent factors, a decision tree was per- ues a significative improvement was found in VAS score
formed to calculate cut-off values and 2 × 2 contingency (7.6 ± 1.3–1.6 ± 0.9, p < 0.001), ODI score at 6 weeks
table for OR. SPSS Statistics 14.0 (SPSS Inc, Richmond, (22.4 ± 6.4) improved to 17.4 ± 6.2 (p < 0.05) at last
CA, USA) was used for data analysis. follow-up and the SF-36 score (62.7 ± 7.1–73.6 ± 6.1,

13
78 European Spine Journal (2023) 32:75–83

Table 1  Patient characteristics replacement of the screw and 8 patients (2.7%) with implant
failure. Five minor complications (1.7%) were observed: 5
Total no. of patients 296
patients had a superficial wound infection with complete
Mean age ± SD, yrs (range) 46.2 ± 19.1 (15–86)
resolution within 2 weeks after surgery.
Mean follow-up ± SD, mos (range) 124.3 ± 26.1 (78–174)
Ten patients required reoperation (overall reoperation rate
Sex
of 3.4%): two patients needed a revision surgery for screw
Female 126 (42.6%)
misplacement, and 8 patients needed a revision surgery for
Male 170 (57.4%)
implant failure (6 patients for screw pull-out, 1 rod disloca-
ASA classification
tion and 1 screw breakage) during follow-up.
I 92 (31.1%)
II 149 (50.3%)
Subgroup analysis of preoperative radiological
III 45 (15.2%)
parameters for implant failure
IV 10 (3.4%)
V 0
In univariate analysis, preoperative MSI (OR = 1.1; 95%
Comorbidity
CI [1.03–1.16]; p = 0.002); and VBHL (OR = 0.95; 95%
Cardiovascular diseases 146 (49.3%)
CI [0.64–1.13]; p = 0.334) were not related to implant fail-
Diabetes mellitus 99 (33.4%)
ure. Additionally, number of fractures (OR = 1.18; 95% CI
Obesity 87 (29.4%)
[1.03–1.16]; p = 0.0038); Cobb angle (OR = 1.4; 95% CI
Respiratory diseases 41 (13.8%)
[1.2–1.6]; p < 0.001); and SI (OR = 1.5; 95% CI [1.2–1.7];
Smokers 96 (32.4%)
p < 0.001).
Traumatic mechanism
The multivariate regression analysis showed that num-
Car/motorbike accident 196 (66.2%)
ber of fractures (OR = 1.1; 95% CI [1.01–1.2]; p = 0.05),
Work activity 40 (13.5%)
preoperative Cobb angle (OR = 1.3; 95% CI [1.12–1.62];
Sport activity 25 (8.4%)
p < 0.001) and SI (OR = 1.5; 95% CI [1.23–1.72]; p < 0.001)
Domestic activity 23 (7.8%)
were independent factors for implant failure. Moreo-
Diving 12 (4.1%)
ver, a decision tree with cut off values showed that Cobb
ASA, American Society of Anesthesiologists angle ≥ 15° (OR = 9.5; 95% CI [2–45.5]; p = 0.005) and
SI ≥ 21° (OR = 39.6; 95% CI [4.9–319.1]; p = 0.0006) pre-
dicts implant failure. The univariate and multivariate analy-
p < 0.05), showed good clinical outcomes with no dis- sis for implant failure are resumed in Table 5.
ability during long-term follow-up. Clinical outcomes are
resumed in Table 3.
Thirty-nine out of the 2048 (1.9%) implanted screws had Discussion
a pedicle breach > 2 mm (grade C or more); however, only
two screws were replaced for occurrence of postoperative Although TL fracture were historically managed using the
radiculopathy. standard open surgery for obtaining segmental fusion, asso-
The Cobb angle (preop 10.2° ± 5.6 vs 4.7° ± 3.6 at fol- ciated to a decompression of the spinal canal when needed,
low-up, p < 0.001) and SI (preop 16.4° ± 7.2 vs 8.7° ± 5.0 at the evolution of surgical techniques and instrumentation
follow-up, p < 0.001) improved significantly after surgery implants have provided alternatives such as percutaneous
and were maintained at follow-up. Similarly, the MSI (preop PPSF. This has demonstrated as effective as standard open
52.2% ± 19.1 vs 3.4% ± 2.8 at follow-up, p < 0.001) and the surgery in terms of clinical-radiological outcomes, while
VBHL ratio (preop 0.55 ± 0.16 vs 0.60 ± 0.19 at follow-up, providing valuable surgical advantages [7, 8]. Accordingly,
p < 0.001) improved significantly after surgery. Radiologi- spine surgeries for TL fractures has been progressively
cal outcomes are resumed in Table 4. An illustrative case is changed with the adoption of minimally invasive surgery
presented in Fig. 1. (MIS) for the management of these patients [27–31].
In cases of burst fractures, indirect decompression can be
Complications and reoperation rate obtained with the ligamentotaxis, consisting in the tension
of the posterior longitudinal ligament during distraction, and
No major or minor intra-operative complications were regis- segmental lordorization, promoting the self-repositioning of
tered. Fifteen complications were recorded (overall compli- the dislocated bone fragments which are pushed forward
cation rate of 5.1%). Ten patients had a major complication [32–34]. The additional screws into the fractured vertebra
that needed a revision surgery: 2 patients (0.7%) experienced provide a supplementary hyperlordorizing force, eventu-
a nerve root radiculopathy with complete resolution after ally ameliorating segmental kyphosis correction grade.

13
European Spine Journal (2023) 32:75–83 79

Table 2  Operative characteristics

Nr. (%)

Total no of fractures 368

Total no of screws 2048
Type of fracture (AOSpine classification)
A3 76 (20.7%)
A4 234 (63.6%)
B1 36 (9.8%)
B2 16 (4.3%)
B3 6 (1.6%)
Level of fracture
Thoracic (T4–T9) 108 (32.1%)
Thoracolumbar junction (T10–L2) 179 (53.3%)
Lumbar (L3–L5) 49 (14.6%)
Nr of fractures in a single patient
One 233 (78.7%)
Two 54 (18.2%)
Three 9 (3.1%)
Complications
Minor 5 (1.7%)
Major 10 (3.4%)
Overall complication rate 5.1%
Implant failure
Rod dislocation 1 (0.34%)
Screw breakage 1 (0.34%)
Screw pull-out 6 (2%)
Screw misplacement 39 (1.9%)*
Reoperation rate
Implant failure 8 (2.7%)
Screw misplacement 2 (0.7%)
Overall reoperation rate 3.4%
Nr. (range)

Mean Thoracolumbar AOSpine Injury Score (TL AOSIS) 5 (3–8)

Mean length of surgery ± SD, min (range) 54.1 ± 23.8 (30–150)
Mean length of stay ± SD, days (range) 2.6 ± 3.6 (2–40)
Mean time of postoperative mobilization, ± SD, days (range) 1.5 ± 3.1 (2–30)
Estimated blood loss (EBL) ± SD, mL (range) 46.6 ± 15.2 (30–110)
*
Calculated from 2048 total screws

Furthermore, it seems to increase the bone healing rate in management is able to provide long-term good results, in
mid-term, then reducing chances for segmental kyphosis terms of clinical outcomes.
[19]. The evaluation of radiological parameters has demon-
Our results confirm data from the pertinent literature, in strated that segmental deformity correction, MSI, SI and
terms of good surgical, clinical and radiological outcomes VBHL result as maintained over the follow-up, as matter
after PPSF for TL fractures [9, 10, 13]. In the present series, of proof of segmental stability. Furthermore, short-segment
PPSF provided a significative pain relief after surgery, along instrumentation, including the fractured vertebra and the
with lower disability grade. Furthermore, these clinical and two contiguous ones, revealed to be effective in maintain-
functional improvements were still significative in a long- ing correction even in TL junction (T10-L2), as measured in
term follow-up, as long as 10 years (range 78–174 months). 179 patients (53.3%). No cases of implants failure or adja-
Accordingly, we firstly reported how PPSF in TL fractures cent segment degeneration were reported in these patients

13
80 European Spine Journal (2023) 32:75–83

Table 3  Clinical outcomes Table 4  Radiological outcomes

Mean ± SD Mean ± SD

Visual Analogue Scale (VAS) Mid-Sagittal Index (MSI)%

Preoperative 7.6 ± 1.3 Preoperative 52.2 ± 19.1
Postoperative (6 weeks) 2.5 ± 1.9 Postoperative (6 weeks) 36.5 ± 14.3
Follow-up at 1 years 2.2 ± 0.8 Follow-up at 1 years 6.1 ± 4.7
Follow-up at 6 years 1.9 ± 1.0 Follow-up at 6 years 3.5 ± 2.9
Last follow-up 1.6 ± 0.9 Last follow-up 3.4 ± 2.8
p value (pre vs follow-up) < 0.001 p value (pre vs fu) < 0.001
Oswestry Disability Index (ODI) Cobb’s Angle°
Preoperative N/A Preoperative 10.2 ± 5.6
Postoperative (6 weeks) 22.4 ± 6.4 Postoperative (6 weeks) 5.4 ± 4.1
Follow-up at 1 years 16.2 ± 5.6 Follow-up at 1 years 4.6 ± 3.5
Follow-up at 6 years 17.7 ± 5.8 FOLLOW-up at 6 years 4.2 ± 3.8
Last follow-up 17.4 ± 6.2 Last follow-up 4.7 ± 3.6
p value (post vs follow-up) < 0.001 p value (pre vs fu) < 0.001
SF-36 (physical and mental) Sagittal Index (SI)°
Preoperative N/A Preoperative 16.4 ± 9.2
Postoperative (6 weeks) 62.7 ± 7.1 Postoperative (6 weeks) 10.4 ± 6.1
Follow-up at 1 years 78.4 ± 6.6 Follow-up at 1 years 9.7 ± 5.5
Follow-up at 6 years 75.3 ± 5.9 Follow-up at 6 years 8.0 ± 5.2
Last follow-up 73.6 ± 6.1 Last follow-up 8.7 ± 5.0
p value (post vs follow-up) < 0.001 p value (pre vs fu) < 0.001
Vertebral body height loss (VBHL) ratio
Preoperative 0.55 ± 0.16
and an anterior approach does not seem justified. A recent Postoperative (6 weeks) 0.58 ± 0.18
meta-analysis confirms these results showing no difference Follow-up at 1 years 0.59 ± 0.20
radiological and functional outcomes between the anterior Follow-up at 6 years 0.64 ± 0.21
and posterior approaches with longer duration and estimated Last follow-up 0.60 ± 0.19
blood loss in the anterior one [35]. p value (pre vs fu) < 0.001
A planned implant removal is not systematically adopted
in our institution. The removal is agreed with sportive
agonistic patients only, while second surgeries for remov- In the over twenty years experience on this, we have
ing the implants are reserved to those patients complain- developed an internat protocol on the management of neu-
ing for the instrumentation and were not included in this rologically intact patients with TL fractures and invasion of
study. Nowadays, there exists ongoing debate if patients the spinal canal in which we do never perform decompres-
with posterior fixation with or without arthrodesis of thora- sion, while obtaining the indirect decompression using liga-
columbar fractures should have their implants routinely mentotaxis as described above. A recent systematic review
removed on follow-up [36–38]. However, in this cohort, strengthened that surgical maneuvers to promote segmental
the implant retention did not lead to disability and lower fusion do not improve clinical or radiological outcomes, but
functional outcomes during the 10 years of follow-up, prob- they are associated with increased surgical time and higher
ably due to the short-segment instrumentation used in most intraoperative bleeding [39].
patients. Moreover, the multivariate logistic regression on In terms of complications, our data showed that PPSF
preoperative radiological parameters showed how number for traumatic TL fractures report a relatively low complica-
of fractures (OR = 1.1), preoperative Cobb angle (OR = 1.3) tion and reoperation rates. The overall complication rate was
and SI (OR = 1.5) were independent risk factors for implant 5.1% (15 patients): 10 patients with major complications
failure. Additionally, a decision tree model showed that nr. that needed a revision surgery and 5 patients with minor
of fractures ≥ 2, preoperative Cobb angle ≥ 15° and sagittal complications (superficial wound infection). The reoperation
index ≥ 21° could predict the subset of patients that could rate was 3.4% (10 patients): in 2 cases one screw was repo-
have an implant failure during follow-up and benefit from a sitioned and in 8 patients we documented implant breakage
different approach. or pull-out.

13
European Spine Journal (2023) 32:75–83 81

Fig. 1  A 34-year-old patient

with a diagnosis of an A4 L2
fracture who had no neurologi-
cal deficits. The preoperative
sagittal (a) and axial (b) CT
scan shows 88% spinal canal
invasion. (b) The postoperative
sagittal (c) and axial (d) CT
scan at the 1-year follow-up
shows percutaneous fixation
with intermediate pedicle
screws. The sagittal (e) and
axial (f) MRI scan at the 5-year
follow-up shows vertebral body
restoration and correction of the
kyphosis angle without spinal
canal invasion. The sagittal
(g) and axial (h) MRI scan at
10-years of follow-up

Table 5  Radiological predictors Univariate analysis Multivariate analysis*

of implant failure
Risk factors OR CI 95% p value OR CI 95% p value

Nr of fractures 1.182 (1.106–1.437) 0.0038 1.102 (1.011–1.412) 0.0495

Type of fracture 1.063 (0.951–1.151) 0.293
Cobb angle 1.405 (1.198–1.647) < 0.001 1.307 (1.124–1.615) < 0.001
SI 1.452 (1.227–1.719) < 0.001 1.452 (1.227–1.719) < 0.001
MSI 1.006 (0.971–1.041) 0.755
VBHL 0.946 (0.641–1.131) 0.334
*
Backward stepwise (conditional) method in binary logistic regression analysis, not significant factors are
eliminated during analysis

To the best of our knowledge, this study retrieved data Limitations of this study
from the largest cohort of patients operated for TL fractures
using PPSF, with the longest follow-up (mean follow-up of There are few limitations to be disclosed. The present
10 years) in the international literature. Our results confirm investigation consists of a single center, retrospec-
that PPSF can be considered as an effective and reproduc- tive study, which has to be considered for a proper data
ible technique for safely managing TL fractures, providing interpretation. Additionally, our study included differ-
valuable clinical and radiological outcomes, and long-term ent types of fractures (type A and B) and the number of
segmental stability. Short segment fixations, involving as few levels instrumented was highly heterogeneous, and this
levels as possible, could be an alternative to longer implants may influence the relevance of our results. At last, the
preserving spinal mobility, thus reducing surgical-mediated subgroup analysis for independent risk factors for implant
restrictions of spinal range of motion. Furthermore, we firstly failure was evaluated only for the preoperative radiologi-
documented how clinical and radiological outcomes are cal parameters. Furthers studies including patients’ char-
maintained over 10 years, with a relatively low complication acteristics, radiological parameters and subgroup analy-
and reoperation rates. Lastly, the multivariate logistic regres- sis for different type of fractures are necessary to better
sion on preoperative radiological parameters showed that nr. evaluate risk factors.
of fractures ≥ 2, preoperative Cobb angle ≥ 15° and sagittal
index ≥ 21° are independent risk factors for implant failure.

13
82 European Spine Journal (2023) 32:75–83

Conclusions 12. Lee JK, Jang JW, Kim TW et al (2013) Percutaneous short-seg-
ment pedicle screw placement without fusion in the treatment of
thoracolumbar burst fractures: is it effective?—comparative study
Percutaneous pedicle screw fixation without arthrodesis with open short-segment pedicle screw fixation with posterolateral
showed excellent long-term clinical and radiological out- fusion. Acta Neurochir (Wien) 155:2305–2312
comes, along with a relatively low complication and reoper- 13. Vanek P, Bradac O, Konopkova R et al (2014) Treatment of thora-
columbar trauma by short-segment percutaneous transpedicular
ation rates, in TL fractures management. Accordingly, PPSF screw instrumentation: prospective comparative study with a
without arthrodesis could be considered as a valuable treat- minimum 2-year follow-up. J Neurosurg Spine 20:150–156
ment for neurologically intact patients with TL fractures. 14. Chi JH, Eichholz KM, Anderson PA et al (2019) Congress of
Additionally, number of fractures ≥ 2, Cobb angle ≥ 15° and neurological surgeons systematic review and evidence-based
guidelines on the evaluation and treatment of patients with thora-
sagittal index ≥ 21° were independent risk factors for implant columbar spine trauma: novel surgical strategies. Neurosurgery
failure. Further randomized comparative studies are neces- 84:E59–E62
sary to confirm these results. 15. Li K, Li Z, Ren X et al (2016) Effect of the percutaneous pedicle
screw fixation at the fractured vertebra on the treatment of thora-
columbar fractures. Int Orthop 40:1103–1110
16. Kapoen C, Liu Y, Bloemers F et al (2020) Pedicle screw fixation
Funding There was no financial support for this study.
of thoracolumbar fractures: conventional short segment versus
short segment with intermediate screws at the fracture level—a
Declarations systematic review and meta-analysis. Eur Spine J 29:2491–2504
17. Norton RP, Milne EL, Kaimrajh DN et al (2014) Biomechanical
Conflict of interest All authors have nothing to declare. analysis of four- versus six-screw constructs for short segment
pedicle screw and rod instrumentation of unstable thoracolumbar
fractures. Spine J 14:1734–1739
18. Baaj AA, Reyes PM, Yaqoobi AS et al (2011) Biomechanical
advantage of the index-level pedicle screw in unstable thora-
columbar junction fractures. J Neurosurg Spine 14:192–197
References 19. Cimatti M, Forcato S, Polli FM et al (2013) Pure percutaneous
pedicle screw fixation without arthrodesis of 32 thoraco-lumbar
1. Hu R, Mustard CA, Burns C (1996) Epidemiology of incident fractures: clinical and radiological outcome with 36-month fol-
spinal fracture in a complete population. Spine (Phila Pa 1976) low-up. Eur Spine J 22:925–932
21:492–499 20. Trungu S, Forcato S, Bruzzaniti P et al (2019) Minimally inva-
2. Holmes JF, Miller PQ, Panacek EA et al (2001) Epidemiology sive surgery for the treatment of traumatic monosegmental thora-
of thoracolumbar spine injury in blunt trauma. Acad Emerg Med columbar burst fractures: clinical and radiologic outcomes of 144
8:866–872 patients with a 6-year follow-up comparing two groups with or
3. Vaccaro AR, Oner C, Kepler CK et al (2013) AOSpine thora- without intermediate screw. Clin Spine Surg 32:E171–E176
columbar spine injury classification system: fracture description, 21. Kepler CK, Vaccaro AR, Schroeder GD et al (2016) The thora-
neurological status, and key modifiers. Spine 38:2028–2037 columbar AOSpine injury score. Global Spine J 6:329–334
4. Urrutia J, Zamora T, Yurac R et al (2015) An independent inter- 22. Glassman SD, Hamill CL, Bridwell KH et al (2007) The impact of
observer reliability and intraobserver reproducibility evaluation perioperative complications on clinical outcome in adult deform-
of the new AOSpine thoracolumbar spine injury classification ity surgery. Spine 32:2764–2770
system. Spine 40:E54–E58 23. Keynan O, Fisher CG, Vaccaro A et al (2006) Radiographic
5. Wood KB, Li W, Lebl DS et al (2014) Management of thora- measurement parameters in thoracolumbar fractures: a systematic
columbar spine fractures. Spine J 14:145–164 review and consensus statement of the spine trauma study group.
6. Dai LY, Jiang SD, Wang XY et al (2007) A review of the manage- Spine 31:E156–E165
ment of thoracolumbar burst fractures. Surg Neurol 67:221–231 24. Sadiqi S, Verlaan JJ, Lehr AM et al (2017) Measurement of
7. Phan K, Rao PJ, Mobbs RJ (2015) Percutaneous versus open kyphosis and vertebral body height loss in traumatic spine frac-
pedicle screw fixation for treatment of thoracolumbar fractures: tures: an international study. Eur Spine J 26:1483–1491
systematic review and meta-analysis of comparative studies. Clin 25. Farcy JPC, Glassman SD (1990) Sagittal index in management of
Neurol Neurosurg 135:85–92 thoracolumbar burst fractures. Spine 15(9):958–965
8. McAnany SJ, Overley SC, Kim JS et al (2016) Open versus mini- 26. Gertzbein SD, Robbins SE (1990) Accuracy of pedicular screw
mally invasive fixation techniques for thoracolumbar trauma: a placement in vivo. Spine 15:11–14
meta-analysis. Global Spine J 29:186–194 27. Jaikumar S, Kim DH, Kam AC (2002) History of minimally inva-
9. Charles YP, Zairi F, Vincent C et al (2012) Minimally invasive sive spine surgery. Neurosurgery 51(suppl_2):S2–S1
posterior surgery for thoracolumbar fractures. New trends to 28. Pietrantonio A, Trungu S, Famà I et al (2019) Long-term clini-
decrease muscle damage. Eur J Orthop Surg Traumatol 22:1–7 cal outcomes after bilateral laminotomy or total laminectomy for
10. Wild MH, Glees M, Plieschnegger C et al (2007) Five-year follow- lumbar spinal stenosis: a single-institution experience. Neurosurg
up examination after purely minimally invasive posterior stabi- Focus 46:E2
lization of thoracolumbar fractures: a comparison of minimally 29. Miscusi M, Trungu S, Forcato S et al (2018) Long-term clinical
invasive percutaneously and conventionally open treated patients. outcomes and quality of life in elderly patients treated with inter-
Arch Orthop Trauma Surg 127:335–343 spinous devices for lumbar spinal stenosis. J Neurol Surg A Cent
11. Sun XY, Zhang XN, Hai Y (2017) Percutaneous versus traditional Eur Neurosurg 79:139–144
and paraspinal posterior open approaches for treatment of thora- 30. Assaker R (2004) Minimal access spinal technologies: state-
columbar fractures without neurologic deficit: a meta-analysis. of-the-art, indications, and techniques. Joint Bone Spine
Eur Spine J 26:1418–1431 71(6):459–469

13
European Spine Journal (2023) 32:75–83 83

31. Trungu S, Ricciardi L, Forcato S et al (2021) Percutaneous instru- 37. Deckey JE, Bradford DS (2000) Loss of sagittal plane correction
mentation with cement augmentation for traumatic hyperextension after removal of spinal implants. Spine 25(19):2453–2460
thoracic and lumbar fractures in ankylosing spondylitis: a single- 38. Kweh BTS, Tan T, Lee HQ et al (2022) Implant removal ver-
institution experience. Neurosurg Focus 51(4):E8 sus implant retention following posterior surgical stabilization
32. Leferink VJM, Nijboer JMM, Zimmerman KW et al (2003) Burst of thoracolumbar burst fractures: a systematic review and meta-
fractures of the thoracolumbar spine: changes of the spinal canal analysis. Global Spine J 12(4):700–718
during operative treatment and follow-up. Eur Spine J 12:255–260 39. Diniz JM, Botelho RV (2017) Is fusion necessary for thoracolum-
33. Yang H, Shi JH, Ebraheim M et al (2011) Outcome of thoracolum- bar burst fracture treated with spinal fixation? A systematic review
bar burst fractures treated with indirect reduction and fixation and meta-analysis. J Neurosurg Spine 27:584–592
without fusion. Eur Spine J 20:380–386
34. Mueller LA, Mueller LP, Schmidt R et al (2006) The phenomenon Publisher's Note Springer Nature remains neutral with regard to
and efficiency of ligamentotaxis after dorsal stabilization of thora- jurisdictional claims in published maps and institutional affiliations.
columbar burst fractures. Arch Orthop Trauma Surg 126:364–368
35. Tan T, Rutges J, Marion T et al (2019) Anterior versus posterior Springer Nature or its licensor holds exclusive rights to this article under
approach in traumatic thoracolumbar burst fractures deemed for a publishing agreement with the author(s) or other rightsholder(s);
surgical management: systematic review and meta-analysis. J Clin author self-archiving of the accepted manuscript version of this article
Neurosci 70:189–197 is solely governed by the terms of such publishing agreement and
36. Oh HS, Seo HY (2019) Percutaneous pedicle screw fixation in applicable law.
thoracolumbar fractures: comparison of results according to
implant removal time. Clin Orthop Surg 11(3):291–296

13