Incidence and risk factors of recurrence in limb osteomyelitis patients after antibiotic-loaded cement spacer for definitive bone defect treatment

Abstract

Aims:
This study was designed to characterize the recurrence incidence and risk factors of antibiotic-loaded cement spacer (ALCS) for definitive bone defect treatment in limb osteomyelitis.

Methods:
We included adult patients with limb osteomyelitis who received debridement and ALCS insertion into the bone defect as definitive management between 2013 and 2020 in our clinical centre. The follow-up time was at least two years. Data on patients' demographics, clinical characteristics, and infection recurrence were retrospectively collected and analyzed.

Results:
In total, 314 patients with a mean age of 52.1 years (SD 12.1) were enrolled. After a mean of 50 months' (24 to 96) follow-up, 53 (16.9%) patients had infection recurrence including 32 tibiae, ten femora, ten calcanea, and one humerus. Of all patients with recurrence, 30 (9.6%) occurred within one year and 39 (12.4%) within two years. Among them, 41 patients needed reoperation, five received antibiotics treatment only, and seven ultimately required amputations. Following multivariable analysis, we found that patients infected with Gram-negative bacilli were more likely to have a recurrence (odds ratio (OR) 2.38, 95% confidence interval (CI) 1.20 to 6.94; p = 0.046) compared to Staphylococcus aureus; segmental bone defects (OR 5.25, 95% CI 1.80 to 15.26; p = 0.002) and smoking (OR 3.00, 95% CI 1.39 to 6.50; p = 0.005) were also independent risk factors for recurrence after treatment.

Conclusion:
Permanent ALCS might be an alternative strategy for definitive bone defect management in selected osteomyelitis cases. However, the overall high recurrence found suggests that it should be cautiously treated. Additionally, segmental defects, Gram-negative infections, and smoking were associated with an increased risk of infection recurrence.

Full text

VOL. 12, NO. 8, AUGUST 2023 467
Freely available online
Follow us @BoneJointRes
BJR
H. Wu,
D. Sun,
S. Wang,
C. Jia,
J. Shen,
X. Wang,
C. Hou,
Z. Xie,
F. Luo
From Southwest
Hospital, Army Medical
University (Third
Military Medical
University), Chongqing,
China
Correspondence should be sent to
Zhao Xie; email:
xiezhao54981@163.com
doi: 10.1302/2046-3758.128.BJR-
2022-0413.R2
Bone Joint Res 2023;12(8):467–
475.
 INFECTION
Incidence and risk factors of recurrence
in limb osteomyelitis patients after
antibiotic‐loaded cement spacer for
definitive bone defecttreatment
Aims
This study was designed to characterize the recurrence incidence and risk factors of antibiotic-
loaded cement spacer (ALCS) for definitive bone defect treatment in limb osteomyelitis.
Methods
We included adult patients with limb osteomyelitis who received debridement and ALCS
insertion into the bone defect as definitive management between 2013 and 2020 in our
clinical centre. The follow- up time was at least two years. Data on patients’ demographics,
clinical characteristics, and infection recurrence were retrospectively collected and analyzed.
Results
In total, 314patients with a mean age of 52.1 years (SD 12.1) were enrolled. After a mean of
50 months’ (24 to 96) follow- up, 53 (16.9%) patients had infection recurrence including 32
tibiae, ten femora, ten calcanea, and one humerus. Of all patients with recurrence, 30 (9.6%)
occurred within one year and 39 (12.4%) within two years. Among them, 41 patients needed
reoperation, five received antibiotics treatment only, and seven ultimately required amputa-
tions. Following multivariable analysis, we found that patients infected with Gram- negative
bacilli were more likely to have a recurrence (odds ratio (OR) 2.38, 95% confidence interval
(CI) 1.20 to 6.94; p = 0.046) compared to Staphylococcus aureus; segmental bone defects (OR
5.25, 95%CI 1.80 to 15.26; p = 0.002) and smoking (OR 3.00, 95%CI 1.39 to 6.50; p = 0.005)
were also independent risk factors for recurrence after treatment.
Conclusion
Permanent ALCS might be an alternative strategy for definitive bone defect management
in selected osteomyelitis cases. However, the overall high recurrence found suggests that it
should be cautiously treated. Additionally, segmental defects, Gram- negative infections, and
smoking were associated with an increased risk of infection recurrence.
Cite this article: Bone Joint Res 2023;12(8):467–475.
Keywords: Osteomyelitis, Risk factors, Recurrence, Definitive treatment, ALCS
Article focus
 Antibiotic- loaded cement spacer (ALCS)
is commonly used but not always applied
for permanent implantation in manage-
ment of osteomyelitis, of which the
overall recurrence and relative contrain-
dications are unknown.
 This study focused on the incidence of
recurrence and the associated risk factors
of permanent ALCS in treating limb
osteomyelitis.
Key messages
 The data show that ALCS can serve as an
alternative method for definitive treat-
ment in some limb osteomyelitis patients.
However, the overall recurrence in this
study suggests that it should be carefully
treated.

BONE & JOINT RESEARCH
H. WU, D. SUN, S. WANG, C. JIA, J. SHEN, X. WANG, C. HOU, Z. XIE, F. LUO
468
 The study also demonstrated that segmental bone
defect, smoking, and Gram- negative bacilli infection
were independent risk factors and represent relative
contraindications in osteomyelitis management.
Strengths and limitations
 The study enrolled a large series of limb osteomyelitis
patients, and is the first to provide insights into their
prognosis after permanent ALCS treatment and the
risk factors associated with recurrence. The study’s
findings will help in counselling patients, preparing
for treatment preoperatively, and improving
outcomes.
 This was a retrospective study. Furthermore, not all
risk factors could be considered within the confines
of this study.
Introduction
The treatment of limb osteomyelitis is continuously
challenging due to high recurrence, complex complica-
tions, and long- term processes. Antibiotic treatment and
radical debridement of all dead and poorly vascularized
tissues are essential to ensure success.1- 3 Debridement
often creates dead spaces or bone defects, and placing
an antibiotic‐loaded cement spacer (ALCS)4- 6 has become
the most common strategy. ALCS is broadly used to treat
osteomyelitis based on high‐dose local antibiotic delivery
and good dead space management.4,7 Moreover, ALCS
can be an alternative method for definitive bone recon-
struction after initial infection control.1, 8 Compared to
common bone reconstruction options of bone graft9 and
bone transport technique,10 the ALCS reconstruction is
simple, highly reproducible, and may lead to satisfactory
results in some cases.8
Nevertheless, ALCS is not always applied for perma-
nent implantation and often needs removal. When ALCS
is used for long- term implantation to treat osteomyelitis,
researchers have recently raised great concern for possible
exacerbation of infections from microbial colonization of
the cement spacer, and promotion of drug resistance at
subtherapeutic antibiotic levels, eventually resulting in
higher recurrence.11,12 In these cases, the recurrence and
relative contraindications become the most concerning
problems in osteomyelitis patients when ALCS is used
for definitive treatment. However, reports have been
published only for the long- term use of ALCS in peri-
prosthetic joint infection (PJI) after joint arthroplasty. To
our knowledge, few studies have indicated that ALCS
is a definitive reconstruction method for limb osteomy-
elitis.1,8 Furthermore, these studies were only case reports
or small case series, and information on the incidence
and risk factors of infection recurrence remains limited.
Herein, we retrospectively reported a large series
of osteomyelitis patients who underwent an ALCS for
definitive treatment after debridement and antimicrobial
therapy. We aimed to identify the incidence of recur-
rence after ALCS for definitive bone defect treatment
in limb osteomyelitis, and further assess the risk factors
associated with infection recurrence. Our findings may
help clinicians to appropriately select patients suitable
for this treatment and the preoperative preparation to
improve outcomes.
Methods
Study design. We performed an eight- year observational
study by reviewing our clinical centre data. All adult pa-
tients treated for limb osteomyelitis between 2013 and
2020 were identified. The Department of Orthopaedics
at Southwest Hospital is a tertiary, high- volume, level-
one bone infection treatment centre, with approximately
300 osteomyelitis patients treated yearly. The inclusion
criteria were: limb osteomyelitis (either haematogenous
or post- traumatic);9 remaining bone defects and ALCS
insertion after debridement; patients over 18 years of
age; medically unfit or refused revision surgery,1,8 and
selected ALCS for definitive bone reconstruction after
infection control; and follow- up time≥ 24months. The
exclusion criteria were: infections without curative treat-
ment (palliative care); incomplete medical records and
follow- up data; and patients with malignant disease.
Patients included in the analysis were classified as either
cured or had recurrence based on follow- up information.
The medical ethics committee of Southwest Hospital ap-
proved this retrospective investigation.
Treatment protocol. Orthopaedic trauma and bone infec-
tion specialists were responsible for the treatment pro-
tocol. A multidisciplinary collaboration,10 including plas-
tic surgeons, microbiologists, pathologists, and clinical
pharmacologists, was established in the Department of
Orthopaedics at Southwest Hospital. All patients received
debridement, standard antimicrobial treatment, and
staged surgical protocol to treat the infection (Figure1).
First, a radical debridement was performed, and all dead
and poorly vascularized tissues were operatively re-
moved. Then, ALCS (500mg gentamicin per 40 mg of
poly(methyl methacrylate) (PMMA) powder mixed with
5 g vancomycin powder)4,13 was used for dead space
management. Cases with instability were reinforced with
additional fixation,14 and the plastic surgeon (JS) per-
formed flap surgery when necessary. All patients were
empirically treated with systemic broad- spectrum anti-
biotics when the microbial cultures were sent for analy-
sis,15 then tailored to the pathogen responsible once the
culture results arrived. Sensitive antibiotics were intrave-
nously administered for two weeks, then orally for four
weeks after discharge.
At least six to eight weeks after debridement, when
an initial clinical infection control was obtained, patients
were considered for definitive bone reconstruction. The
initial clinical infection control was defined as no clin-
ical signs of infection, normalization of white blood cell
(WBC) counts and CRP, and adequate healing of the soft-
tissue.1,6,16 If there were clinical signs of infection, a second
debridement was performed. In the ALCS, PMMA first
served as an antibiotic carrier and dead space manage-
ment tool, and ultimately as a form of definitive bone

VOL. 12, NO. 8, AUGUST 2023
INCIDENCE AND RISK FACTORS OF RECURRENCE IN LIMB OSTEOMYELITIS PATIENTS 469
reconstruction. ALCS was selected for definitive treat-
ment based on the patient’s willingness8 and clinicians’
experience.1 These patients were followed up without
any further treatment (Figures2 and 3).
Definition and observation measures. The definition of
osteomyelitis recurrence followed the consensus criteria
on fracture- related infections (FRIs) of the European Bone
and Joint Infection Society (EBJIS), confirmed by the pres-
ence of at least one of the following findings: 1) fistula,
sinus, or wound breakdown; 2) purulent drainage from
the wound or presence of pus during surgery; 3) path-
ogens identified by culture from at least two separate
deep tissue/implant specimens; and 4) the presence of
Fig. 1
Staged treatment protocol for limb osteomyelitis remaining bone defects. ALCS, antibiotic‐loaded cement spacer; PMMA, poly(methyl methacrylate).
Fig. 2
Successful management of a 72- year- old male osteomyelitis patient in distal femur with an antibiotic‐loaded cement spacer (ALCS) for definitive treatment. a)
Photograph presenting a sinus in distal, lateral side of the thigh. b) Anteroposterior and lateral radiographs before debridement. c) Radiographs of three, 12,
and 24months after an ALCS for definitive treatment, where the implants and cement spacer were in place with no excessive bone loss.

BONE & JOINT RESEARCH
H. WU, D. SUN, S. WANG, C. JIA, J. SHEN, X. WANG, C. HOU, Z. XIE, F. LUO
470
microorganisms in deep tissues during an operative inter-
vention, confirmed by histopathological examination.3 ,17
The main outcome measure was infection recurrence
after treatment. The time of recurrence and related
complications were also collected. This information
was finally converted into the population’s recurrence
and infection cure rates. Infection ‘cure’ was defined as
infection- free after treatment in the two- year follow- up.14
A special fellowship- trained clinical researcher (SW)
prospectively conducted the follow- up after treatment,
mainly via in- person outpatient visits. Patients’ functional
activities and complications were also recorded. Then,
the follow- up information was retrospectively analyzed.
Patient demographic information was extracted
from the electronic medical records, including sex, age,
smoking history, comorbidities (e.g. coronary heart
disease, hypertension, diabetes mellitus, and anaemia),
Cierny- Mader (C- M) host type,14 previous operation,
infection aetiology, infection site, and infection duration.
Other treatment- related variables, such as bone defect
feature, identified microbiological results, fixation type,
and additional flap coverage, were also collected. These
variables were used as exposure factors for the risk factor
analysis.
Statistical analysis. Data were analyzed using SPSS v22.0
(IBM, USA). The Kolmogorov- Smirnov test was used to
verify normalization. Independent- samples t- tests (con-
tinuous variables) and Fisher’s exact tests (categorical var-
iables) were used to compare patient characteristics by
osteomyelitis recurrence. Statistically significant variables
(p < 0.05) in the single- factor analysis were used in the
multivariable logistic regression model. Binary logistic
regression analyses and the Cox regression model were
used to assess the independent association between
these potential factors and infection recurrence.
Results
Study sample and patient characteristics. From 2013 to
2020, 1,839patients with a discharge diagnosis of os-
teomyelitis were identified in our initial search. Then,
1,253 osteomyelitis patients were excluded because
they received bone grafts or bone transport techniques
for reconstruction. A total of 246patients were exclud-
ed because they were under 18 years of age or did not
have limb involvement. Overall, 26 patients who had less
than two years of follow- up were also excluded. Finally,
314patients with limb osteomyelitis fulfilled our criteria
and were included in this study. The demographics and
clinical characteristics of the study sample are presented
in TableI.
Among the study sample, 236 patients (75.2%) were
male, and the mean age was 52.1years (SD 12.1). The
infection sites included 160 in tibia (48 proximal, 47
shaft, 65 distal), 89 in femur (14 proximal, 30 shaft, 45
distal), 17 in upper limbs (three ulna, five radius, and
nine humerus), and 48 in the calcaneus. According to
the C- M classification of osteomyelitis, only 37 (11.8%)
patients were Host A, and the remaining were Host B due
to local or systemic diseases, while the majority (92.1%)
were C- M type Ⅲ among all long bone infections (n =
266). Bacteria were initially isolated in 78.7% (247/314)
of patients, while the remaining 21.3% (67/314) were
culture- negative. Staphylococcus aureus was the main
agent in 40.1% (126/314), methicillin- resistant S. aureus
(MRSA) in 13% (13/314), Gram- negative bacilli were
also common with 29.9% (94/314) (e.g. Pseudomonas
Fig. 3
Recurrent osteomyelitis patient (56 years old, male) with Pseudomonas aeruginosa infection in tibia shaft, who chose an antibiotic‐loaded cement spacer
(ALCS) for definitive treatment. a) Photograph presenting a sinus in midshaft, anterior side of the tibia. b) Anteroposterior (AP) and lateral radiographs before
and after operation of debridement and placing an ALCS for dead space. c) Radiographs (AP and lateral) and photograph showing a sign of recurrence at
13months, where the sinus occured in a previous site.

VOL. 12, NO. 8, AUGUST 2023
INCIDENCE AND RISK FACTORS OF RECURRENCE IN LIMB OSTEOMYELITIS PATIENTS 471
aeruginosa, Escherichia coli, and Enterobacter cloacae),
and 25.2% (79/314) were polymicrobial infections. Most
(88.2%; 277/314) patients had debridement and antimi-
crobial therapy before being admitted to our clinic.
Furthermore, 285 (90.8%) patients received single-
stage debridement and antimicrobial treatments, and
the remaining 29 (9.2%) needed a second debridement
due to incomplete initial infection control in the staged
evaluation. Meanwhile, 43 (13.7%) patients performed
additional flap coverage: 37 in the tibia (25 free and
12 rotational flaps) and six in the calcaneus (rotational
flaps). All patients could walk independently after
wound healing and were satisfied with the treatment.
Most (210/314) patients were medically unfit for other
bone reconstructions based on the clinician’s experience.
Some (75/314) patients had low physical demand, were
satisfied with the current treatment, and refused further
surgery, while the remaining (29/314) patients refused
due to socioeconomic reasons. Thus, they chose ALCS for
definitive bone reconstruction.
Characteristics of recurrence. The mean follow- up time
was 50 months (24 to 96) and 53 (16.9%) patients had a
recurrence, including 32 tibiae, ten femora, ten calcanea,
and one humerus. The overall recurrence decreased over
time; most recurrences (n = 30) occurred in the first year,
and 39 within two years. The two- year infection cure rate
was 87.6%. The recurrence rate for S. aureus was 19.0%
(24/126), and for MRSA was 23.1% (3/13). Gram- negative
bacilli were more common in patients with recurrence
(22/94), in which the main pathogens were P. aerug-
inosa, E. coli, and E. cloacae with 21.2% (7/33), 25.0%
(6/24), and 41.2% (7/17) recurrence rates, respectively.
The recurrence of Gram- negative bacilli occurred mainly
during early follow- up (TableII).
Among the 53patients with recurrence, 41 accepted
re- debridement and insertion of a new ALCS to achieve
infection resolution. The mean number of procedures
before infection resolution was 2.5 (2 to 5). Moreover,
five recurrent femur patients had infection resolution
by antibiotics treatment only. The remaining seven
patients with infection recurrence ultimately required
amputations: five in the tibia, one in the femur, and
one in the calcaneus. These patients could not tolerate
any treatment burden and finally chose an amputation
after repeated failure. Except for those with amputation,
patients had the cement spacer in place, and none had
radiological evidence of excessive bone loss or suered
pain daily during the follow- up (Figures2 and 3).
Risk factors of recurrence. The single- factor analysis was
first conducted to evaluate the association between po-
tential factors and infection recurrence. Since infection
cure is generally defined as “an infection- free at the two-
year follow- up”, we mainly evaluated factors aecting the
recurrence within this period. Segmental bone defects,
Table I. Characteristics in two cohorts with or without recurrence within two years of follow- up.
Characteristic Total (n = 314) No recurrence (n = 275) Recurrence (n = 39) p- value
Mean age, yrs (SD) 53.1 (12.1) 53.0 (12.3) 53.7 (10.7) 0.736*
Male, n (%) 236 (75.2) 204 (74.2) 32 (82.2) 0.329†
Smoker, n (%) 142 (45.2) 116 (42.2) 26 (66.7) 0.006†
Diabetes, n (%) 35 (11.1) 26 (9.5) 9 (23.1) 0.025†
Systemic disease, n (%) 90 (28.7) 74 (26.9) 16 (41.0) 0.088†
Aetiology of traumatic, n (%) 201 (64) 173 (62.9) 28 (71.8) 0.373†
Infection site, n (%)
Femur (ref) 89 (28.3) 82 (29.8) 7 (17.9)
Tibia 160 (51.0) 138 (50.2) 22 (56.4) 0.217†
Calcaneus 48 (15.3) 39 (14.2) 9 (23.1) 0.092†
Upper limb 17 (5.4) 16 (5.8) 1 (2.6) 0.777†
Sinus, n (%) 279 (88.9) 242 (88.0) 37 (94.9) 0.280†
Infection duration, n (%)
≤ 3 mths (ref) 57 (18.2) 53 (19.3) 4 (10.3)
3 to 24 mths 97 (30.9) 83 (30.5) 14 (33.3) 0.414†
≥ 24mths 160 (51) 136 (50.0) 24 (57.1) 0.169†
Cierny- Mader host B, n (%) 276 (87.9) 238 (86.5) 38 (97.4) 0.064†
Prior debridement (≥ 2), n (%) 195 (62.1) 163 (59.3) 32 (82.1) 0.007†
Second debridement, n (%) 29 (9.2) 27 (9.8) 2 (5.1) 0.552†
Segmental bone defects, n (%) 21 (6.7) 12 (4.3) 9 (21.4) 0.001†
Polymicrobial, n (%) 79 (25.2) 67 (24.4) 12 (30.8) 0.431†
Staphylococcus aureus, n (%) 126 (40.1) 112 (40.7) 14 (35.9) 0.605†
Gram- negative bacilli, n (%) 94 (29.9) 72 (26.2) 22 (56.4) ＜0.001†
Flap coverage, n (%) 43 (13.7) 33 (12.0) 10 (25.6) 0.042†
Internal fixation, n (%) 139 (44.3) 119 (43.3) 20 (51.3) 0.391†
*Independent- samples t- test.
†Fisher’s exact test.
ref, reference; SD, standard deviation.

BONE & JOINT RESEARCH
H. WU, D. SUN, S. WANG, C. JIA, J. SHEN, X. WANG, C. HOU, Z. XIE, F. LUO
472
smoking, diabetes, Gram- negative bacilli infections, pre-
vious multiple (two or more) debridements, and flap
coverage were significantly related to recurrence (TableI)
and were further included in the multivariable binary re-
gression analysis (TableIII). In the multivariable analysis,
patients infected with Gram- negative bacilli were more
likely to have a recurrence (odds ratio (OR) 2.38, 95%
confidence interval (CI) 1.20 to 6.94; p = 0.046, logistic
regression) compared to S. aureus, while segmental bone
defects (OR 5.25, 95%CI 1.80 to 15.26; p = 0.002, logistic
regression) and smoking (OR 3.00, 95%CI 1.39 to 6.50;
p = 0.005, LR test) were also independent risk factors
for recurrence after treatment. The overall infection- free
survival time related to these risk factors significantly de-
creased as demonstrated by Cox regression model anal-
ysis, in considering the total follow- up period (Figure4).
Discussion
One of the central principles of surgical treatment for
osteomyelitis is to ensure that the dead space or bone
defects are treated correctly after debridement. The
staged procedure, also known as the Belfast technique,
was first described by McNally et al18 and has become the
standard for osteomyelitis management. This two- stage
protocol decreases surgical morbidity by separating
infection treatment from restoring form and function.
Applying an ALCS has become the most common strategy
to eliminate dead space and delay bone reconstruc-
tion.1,4,7,12 Previous studies have reported that a cement
spacer is an alternative method for definitive bone recon-
struction to treat limb osteomyelitis.1,8 However, ALCS
is not always suitable for permanent implantation, and
the information is still limited.1,8 Herein, we have supple-
mented the literature on this topic.
We found a 16.9% overall recurrence rate in osteo-
myelitis patients (n = 314) with an ALCS for definitive
treatment. Previous studies have found more promising
results but only included a few cases.1,8 For example,
Cierny and DiPasquale1 first reported a case series of adult
osteomyelitis patients (n = 22) treated with a permanent
ALCS, and only a 12% recurrence rate occurred within
two years of follow- up. Qiu et al8 also reported eight
cases of osteomyelitis treated with this method, and all
patients achieved an infection cure. On the other hand,
a recurrence rate of 0% to 12.26% has been reported for
osteomyelitis patients treated with other, more advanced
techniques.3,6,19–21 Yalikun et al19 reported 149 cases of
tibial osteomyelitis treated with the bone transport
Table II. Initial microbiological characteristics in all patients (n = 314) and recurrence (n = 53).
Bacteria type All patients, n (%) Recurrence
Within 1yr, n (%) Within 2yrs, n (%) Total, n (%)
Total, n (%) 314 (100) 30 (9.6) 39 (12.4) 53 (16.9)
Staphylococcus aureus, n (%) 126 (40.1) 11 (8.7) 14 (11.1) 24 (19.0)
MRSA, n (%) 13 (4.1) 1 (7.7) 2 (15.4) 3 (23.1)
GNB, n (%)
Pseudomonas aeruginosa 33 (10.5) 5 (15.2) 5 (15.2) 7 (21.2)
Escherichia coli 24 (8.0) 5 (20.8) 5 (20.8) 6 (25.0)
Enterobacter cloacae 17 (5.4) 4 (23.5) 5 (29.4) 7 (41.2)
Other GNB* 28 (7.0) 5 (17.8) 8 (28.6) 9 (32.1)
Staphylococcus epidermidis, n (%) 25 (8.0) 2 (8.0) 2 (8.0) 2 (16.0)
Polymicrobial, n (%) 79 (25.2) 10 (12.7) 12 (15.2) 16 (20.3)
None, n (%) 67 (21.3) 2 (3.0) 3 (4.5) 5 (7.5)
*Other GNB: Acinetobacter baumannii, Klebsiella pneumoniae, Proteus, Serratia marcescens.
GNB, Gram- negative bacilli; MRSA, methicillin- resistant Staphylococcus aureus.
Table III. Multivariable analysis for risk factors of recurrence (n = 39) within two years of follow- up.
Variable OR 95%CI p- value
Flap coverage 1.15 0.44 to 2.99 0.777
Diabetes: yes vs no (ref) 2.24 0.87 to 5.76 0.094
Prior debridement (n): ≥ 2 vs ≤ 1 (ref) 2.40 0.96 to 6.00 0.062
Bacterials: Staphylococcus aureus (ref)
Gram- negative bacilli 2.38 1.20 to 6.94 0.046
Others 0.51 0.17 to 1.51 0.223
Smoker: yes vs no (ref) 3.00 1.39 to 6.50 0.005
Bone defects: segmental versus cavity (ref) 5.25 1.80 to 15.26 0.002
Model explained variance (R2) was 24.2%.
CI, confidence interval; OR, odds ratio; ref, reference.

VOL. 12, NO. 8, AUGUST 2023
INCIDENCE AND RISK FACTORS OF RECURRENCE IN LIMB OSTEOMYELITIS PATIENTS 473
technique, and an 11.4% recurrence rate was achieved
with a mean follow- up of 37.5months. Similarly, Wang et
al21 recently showed that the recurrence rate of osteomy-
elitis (n = 424) was 12.26%. These patients were treated
with the induced membrane technique and the mean
follow- up time was 31.6months. Given the likelihood of
recurrence rate (16.9%) in this large series, ALCS should
be cautiously considered as a definitive treatment in limb
osteomyelitis.
Moreover, specific indications for a permanent ALCS
treatment in limb osteomyelitis are not established and
generally depend on clinicians’ experience and patients’
willingness, as in our current research. In 2006, Cierny
and DiPasquale1 first introduced the ALCS for definitive
treatment (also called ‘permanent spacers’) of limb osteo-
myelitis based on the clinician’s experience. Qiu et al8
also reported that patients who chose ALCS for definitive
treatment were either medically unfit (n = 2) or refused
further revision (n = 6) in stage management of osteomy-
elitis. In this study, most patients were medically unfit for
other bone reconstructions, mainly because appropriate
bone graft materials were unavailable in most situations
or patients were unsuitable to perform a revision surgery
with high- risk factors (e.g. old age) based on clinicians’
experience. Additionally, some patients had low phys-
ical demand, were satisfied with a permanent spacer
treatment, and refused further surgery. The remaining
patients could not aord the subsequent treatment
due to the high economic burden.22 A cost of $17,000
to $150,000 per patient has been reported,23,24 which is
much worse in countries with limited resources.
Besides the above clinical setting, we found possible
contraindications of this method based on the risk factor
analysis. We identified several independent risk factors for
osteomyelitis recurrence after ALCS for definitive treat-
ment. As a successful infection cure is generally defined
as “an infection- free at the two- year follow- up”,17 we
mainly evaluated the recurrence of patients within this
period. Segmental defects, smoking, and Gram- negative
infections were independent risk factors for long- term
success. First, smoking has been reported as an important
risk factor for various bone infections,25,26 aecting
patients’ prognosis by damaging their immune status.
Thus, educating patients to quit smoking is crucial.
Furthermore, other host- related factors, including
multiple (two or more) previous debridements, diabetes,
and flap coverage, were included in our multivariable anal-
ysis, all known complicating factors for osteomyelitis.14,21
Additionally, segmental bone defects due to osteomy-
elitis (often due to C- M type IV lesions)16,27 are generally
more complex, and the treatment is challenging. Previ-
ously, Cierny14 reported permanent spacers with addi-
tional fixation in 8% (25/314) of segmental osseous defect
patients due to osteomyelitis; however, no treatment
result was available. When ALCS was applied for defini-
tive treatment in this study, a high recurrence rate (42%)
was detected in segmental bone defects, comprising
an independent risk factor. The cement may not have
integrated well with the bone, which would explain
these poor results. There are inevitable frequent interac-
tions between bone cement and bone under segmental
defects, similar to prosthesis,25 in which micro- friction
between the prosthesis and bone surface is considered
a risk factor for joint infection. Similarly, Belay et al28 also
reported a high recurrence rate (20.7%) with a perma-
nent articulating spacer to manage chronic PJI. Based
on these results, segmental bone defects due to osteo-
myelitis seem to be a relative contraindication for using
ALCS as the definitive treatment, which requires further
investigation.
Furthermore, we hypothesized that microbiological
aspects may have specific prognostic implications in limb
osteomyelitis patients after ALCS treatment. Within the
two- year follow- up period, Gram- negative bacilli (e.g.
P. aeruginosa, E. coli, and E. cloacae) had a higher recur-
rence rate than other bacteria (23.4 vs 7.7%). Specifically,
E. cloacae had the highest recurrence rate (29.4%). Osteo-
myelitis caused by Gram- negative bacilli was associated
Fig. 4
Survival function curves using a Cox regression model for risk factors of: a) Gram- negative bacilli (risk ratio (RR) 2.24; p < 0.004); b) smoking (RR 1.85; p <
0.031); and c) segmental bone defect types (RR 3.39; p = 0.001).

BONE & JOINT RESEARCH
H. WU, D. SUN, S. WANG, C. JIA, J. SHEN, X. WANG, C. HOU, Z. XIE, F. LUO
474
with more than a twofold increase in recurrence (OR 4.06;
p = 0.046, multivariable logistic regression) compared to
S. aureus. However, the relationship between treatment
outcomes in osteomyelitis and bacterial type is contra-
dictory. Cierny and DiPasquale1 stated that there is no
relationship between them by reviewing a large treat-
ment cohort (n = 1,966). Other reported series have indi-
cated that Gram- negative osteomyelitis is a therapeutic
challenge,29 especially infection caused by P. aeruginosa
showing a marker for poor prognosis in its manage-
ment.20,21 Here, insucient debridement and inadequate
antimicrobial therapy against Gram- negative bacteria
may be responsible for the poor outcomes. Since most
patients had an early- stage recurrence (within one year),
the added high- dose vancomycin was mainly used for
Gram- positive bacteria. These bacteria might also secrete
special virulence factors,29,30 compromise host immunity,
and increase the resistance to local gentamicin. Besides,
67 (21.3%) patients in this series were culture- negative,
possibly due to antibiotic exposure before sampling,
because 277 (88.2%) patients had received antimicro-
bial therapy before being admitted to our clinic. Notably,
negative cultures were not associated with the patient’s
prognosis in previous studies.31 These results suggest
that ALCS as a definitive treatment may not be suitable
for some refractory bacteria, while radical debridement,
more prolonged courses, and/or intense regimens (or
combinations of antibiotic agents) are recommended.
Although ALCS is most commonly used to manage
dead space and provide high‐dose local antibiotic treat-
ment, many researchers have raised concerns about its
long- term implantation.11,12 The likelihood of infection
recurrence in the current study might support these
concerns. First, the long- term presence of cement can
promote the adherence of microbes and biofilm forma-
tion.29,32 In this study, the higher late recurrence rate
occurred in series caused by S. aureus over time, the
most common biofilm bacteria, which might oer some
evidence. Second, ALCS can only maintain an eec-
tive bactericidal concentration within a few weeks after
implantation, which raises concerns about promoting
bacterial resistance at subtherapeutic antibiotic levels.
Some in vitro studies33,34 have found drug- resistant S.
aureus and Staphylococcus epidermidis on the surface of
bone cement containing gentamicin, tobramycin, and
vancomycin. Despite not being addressed here, the total
recurrence rate indicated the potential risk of promoting
microbe adherence and drug resistance, which need
further investigation. Some scholars recommend that the
ALCS should be removed, if possible, within a few weeks
after local treatment.11,12,33
However, our current study also has some limitations.
First, since this was a retrospective study, randomiza-
tion was impossible, and the treatment of osteomyelitis
heavily relied on the surgeon’s clinical experience. Addi-
tionally, not all risk factors could be considered. Never-
theless, there are also several strengths to this study: we
provided many patients treated with standard techniques
from a high- volume, level- one bone infection treatment
centre. In addition, because this series of patients were
treated by the same team of specialists, we decreased
the impact of subjective variables (e.g. debridement)
that may confound the outcomes. In the multivariable
analysis, adjustments were made for diabetes, host type,
and multiple prior debridements, all known compli-
cating factors for osteomyelitis. Furthermore, the length
of patient follow- up demonstrated that the probability
of recurrence decreases over time. We supplemented
the information in the literature; more evidence- based
studies with longer follow- up durations are needed on
this topic.
In conclusion, we enrolled a large series of limb
osteomyelitis patients and provided insights into their
prognosis after ALCS for definitive treatment and the
risk factors associated with recurrence. The likelihood
of recurrence suggests that ALCS should be cautiously
considered as a definitive treatment, although it can be an
alternative option in some selected osteomyelitis cases.
The data also indicate that segmental defects, Gram-
negative infections, and smoking were associated with
an increased risk of infection recurrence. These patients
need to be carefully treated to achieve long- term success.
References
1. Cierny G, DiPasquale D. Treatment of chronic infection. J Am Acad Orthop Surg.
2006;14(10 Spec No):S105–10.
2. Walter N, Rupp M, Baertl S, Alt V. The role of multidisciplinary teams in
musculoskeletal infection. Bone Joint Res. 2022;11(1):6–7.
3. Baertl S, Metsemakers W- J, Morgenstern M, et al. Fracture- related infection.
Bone Joint Res. 2021;10(6):351–353.
4. Metsemakers W- J, Fragomen AT, Moriarty TF, et al. Evidence- based
recommendations for local antimicrobial strategies and dead space management in
fracture- related infection. J Orthop Trauma. 2020;34(1):18–29.
5. McNally MA, Ferguson JY, Scarborough M, Ramsden A, Stubbs DA, Atkins
BL. Mid- to long- term results of single- stage surgery for patients with chronic
osteomyelitis using a bioabsorbable gentamicin- loaded ceramic carrier. Bone Joint
J. 2022;104- B(9):1095–1100.
6. Shen J, Sun D, Fu J, Wang S, Wang X, Xie Z. Management of surgical site
infection post- open reduction and internal fixation for tibial plateau fractures. Bone
Joint Res. 2021;10(7):380–387.
7. Zalavras CG, Patzakis MJ, Holtom P. Local antibiotic therapy in the treatment of
open fractures and osteomyelitis. Clin Orthop Relat Res. 2004;427(427):86–93.
8. Qiu XS, Zheng X, Shi H, etal. Antibiotic- impregnated cement spacer as definitive
management for osteomyelitis. BMC Musculoskelet Disord. 2015;16:254.
9. Shen J, Sun D, Yu S, et al. Radiological and clinical outcomes using induced
membrane technique combined with bone marrow concentrate in the treatment of
chronic osteomyelitis of immature patients. Bone Joint Res. 2021;10(1):31–40.
10. Tsang S- TJ, Ferreira N, Simpson AHRW. The reconstruction of critical bone loss:
the holy grail of orthopaedics. Bone Joint Res. 2022;11(6):409–412.
11. Schwarz EM, McLaren AC, Sculco TP, etal. Adjuvant antibiotic‐loaded bone
cement: Concerns with current use and research to make it work. J Orthop Res.
2021;39(2):227–239.
12. Samelis PV, Papagrigorakis E, Sameli E, Mavrogenis A, Savvidou O,
Koulouvaris P. Current concepts on the application, pharmacokinetics and
complications of antibiotic- loaded cement spacers in the treatment of prosthetic joint
infections. Cureus. 2022;14(1):e20968.
13. Wang X, Luo F, Huang K, Xie Z. Induced membrane technique for the treatment of
bone defects due to post- traumatic osteomyelitis. Bone Joint Res. 2016;5(3):101–105.
14. Cierny G. Surgical treatment of osteomyelitis. Plast Reconstr Surg.
2011;127 Suppl 1:190S–204S.

VOL. 12, NO. 8, AUGUST 2023
INCIDENCE AND RISK FACTORS OF RECURRENCE IN LIMB OSTEOMYELITIS PATIENTS 475
Author information:
H. Wu, MD, Professor, Postdoctoral researcher, Department of Orthopaedics, South-
west Hospital, Army Medical University (Third Military Medical University), Chong-
qing, China; Department of Orthopedics, Navy 905th Hospital, Naval Medical Uni-
versity, Shanghai, China.
D. Sun, MD, Professor
S. Wang, MD, Professor
C. Jia, MD, Professor
J. Shen, MD, Professor
X. Wang, MD, Professor
Z. Xie, MD, PhD, Professor
F. Luo, MD, Professor
Department of Orthopaedics, Southwest Hospital, Army Medical University (Third
Military Medical University), Chongqing, China.
C. Hou, MD, Professor, Department of Anatomy, Key Laboratory for Biomechanics
and Tissue Engineering of Chongqing, Army Medical University (Third Military
Medical University), Chongqing, China.
Author contributions:
H. Wu: Investigation, Data curation, Writing – original draft.
D. Sun: Investigation, Data curation, Writing – original draft.
S. Wang: Investigation, Resources, Data curation.
C. Jia: Methodology, Supervision.
J. Shen: Methodology, Investigation.
X. Wang: Methodology, Supervision.
C. Hou: Conceptualization, Formal analysis, Writing – review & editing.
Z. Xie: Conceptualization, Formal analysis, Writing – review & editing.
F. Luo: Conceptualization, Formal analysis, Writing – review & editing.
C. Hou, Z. Xie, and F. Luo contributed equally to this work.
H. Wu, D. Sun, and S. Wang contributed equally to this work.
H. Wu, D. Sun, and S. Wang are joint first authors.
Funding statement:
The authors disclose receipt of the following financial or material support for the re-
search, authorship, and/or publication of this article: this work was supported by the
Chongqing Natural Science Foundation Program (cstc2021jcyj- msxmX0541), the
National Natural Science Foundation Program (No. 82202707, No. 82272515), Major
Military Logistics Research Projects (AWS17J004- 02), and the Technical Innovation
Program (CX2019LC121) in Major Clinical Fields of Army Military Medical University.
ICMJE COI statement:
All the authors declare that there are no conflicts of interest.
Data sharing:
The data that support the findings for this study are available to other researchers
from the corresponding author upon reasonable request.
Acknowledgements:
We thank the patients and all sta who participated in treatment of these patients.
Ethical review statement:
This study was approved by the ethics committee of Southwest Hospital Chongqing,
China (No. KY201878).
Open access funding:
The authors report that they received open access funding for their manuscript from
the Chongqing Natural Science Foundation Program (cstc2021jcyj- msxmX0541),
the National Natural Science Foundation Program (No. 82202707, No. 82272515),
Major Military Logistics Research Projects (AWS17J004- 02), and the Technical Inno-
vation Program (CX2019LC121) in Major Clinical Fields of Army Military Medical Uni-
versity. The financial support had no impact on the outcomes of this study.
© 2023 Author(s) et al. This is an open- access ar ticle distributed under the terms of
the Creative Commons Attribution Non- Commercial No Derivatives (CC BY- NC- ND 4.0)
licence, which permits the copying and redistribution of the work only, and provided
the original author and source are credited. See https://creativecommons.org/licenses/
by-nc-nd/4.0/
15. Stengel D, Bauwens K, Sehouli J, Ekkernkamp A, Porzsolt F. Systematic
review and meta- analysis of antibiotic therapy for bone and joint infections. Lancet
Infect Dis. 2001;1(3):175–188.
16. Wu H, Shen J, Yu X, etal. Two stage management of Cierny- Mader type IV chronic
osteomyelitis of the long bones. Injury. 2017;48(2):511–518.
17. Metsemakers WJ, Morgenstern M, McNally MA, et al. Fracture- related
infection: A consensus on definition from an international expert group. Injury.
2018;49(3):505–510.
18. McNally MA, Small JO, Tofighi HG, Mollan RA. Two- stage management of
chronic osteomyelitis of the long bones. The Belfast technique. J Bone Joint Surg Br.
1993;75- B(3):375–380.
19. Yalikun A, Yushan M, Li W, Abulaiti A, Yusufu A. Risk factors associated
with infection recurrence of posttraumatic osteomyelitis treated with Ilizarov bone
transport technique- a retrospective study of 149 cases. BMC Musculoskelet Disord.
2021;22(1):573.
20. Tice AD, Hoaglund PA, Shoultz DA. Risk factors and treatment outcomes in
osteomyelitis. J Antimicrob Chemother. 2003;51(5):1261–1268.
21. Wang X, Wang S, Fu J, Sun D, Shen J, Xie Z. Risk factors associated with
recurrence of extremity osteomyelitis treated with the induced membrane technique.
Injury. 2020;51(2):307–311.
22. Walter N, Hierl K, Brochhausen C, Alt V, Rupp M. The epidemiology and direct
healthcare costs of aseptic nonunions in Germany - a descriptive report. Bone Joint
Res. 2022;11(8):541–547.
23. Jiang N, Wu HT, Lin QR, Hu YJ, Yu B. Health care costs of post- traumatic
osteomyelitis in China: Current situation and influencing factors. J Surg Res.
2020;247:356–363.
24. Stanley CM, Rutherford GW, Morshed S, Coughlin RR, Beyeza T. Estimating
the healthcare burden of osteomyelitis in Uganda. Trans R Soc Trop Med Hyg.
2010;104(2):139–142.
25. Lenguerrand E, Whitehouse MR, Beswick AD, etal. Risk factors associated with
revision for prosthetic joint infection following knee replacement: an observational
cohort study from England and Wales. Lancet Infect Dis. 2019;19(6):589–600.
26. Hogan A, Heppert VG, Suda AJ. Osteomyelitis. Arch Orthop Trauma Surg.
2013;133(9):1183–1196.
27. Bezstarosti H, Metsemakers WJ, van Lieshout EMM, etal. Management of
critical- sized bone defects in the treatment of fracture- related infection: a systematic
review and pooled analysis. Arch Orthop Trauma Surg. 2021;141(7):1215–1230.
28. Belay ES, Wixted CM, Kim BI, etal. A permanent articulating spacer versus two-
stage exchange for chronic periprosthetic joint infection: A propensity score- matched
study. J Arthroplasty. 2023;S0883- 5403(23)00057- 8.
29. Carvalho VC, Oliveira PR, Dal- Paz K, Paula AP, Félix Cda S, Lima AL.
Gram- negative osteomyelitis: clinical and microbiological profile. Braz J Infect Dis.
2012;16(1):63–67.
30. Hamad C, Chowdhry M, Sindeldecker D, Bernthal NM, Stoodley P, McPherson
EJ. Adaptive antimicrobial resistance, a description of microbial variants, and their
relevance to periprosthetic joint infection. Bone Joint J. 2022;104- B(5):575–580.
31. Gitajn IL, Heng M, Weaver MJ, Ehrlichman LK, Harris MB. Culture- negative
infection after operative fixation of fractures. J Orthop Trauma. 2016;30(10):538–544.
32. Nana A, Nelson SB, McLaren A, Chen AF. What’s new in musculoskeletal
infection: Update on biofilms. J Bone Joint Surg Am. 2016;98- A(14):1226–1234.
33. Thomes B, Murray P, Bouchier- Hayes D. Development of resistant strains of
Staphylococcus epidermidis on gentamicin- loaded bone cement in vivo. J Bone Joint
Surg Br. 2002;84- B(5):758–760.
34. Cai Y, Huang C, Chen X, etal. The role of Staphylococcus aureus small colony
variants in intraosseous invasion and colonization in periprosthetic joint infection.
Bone Joint Res. 2022;11(12):843–853.