1473-2262
Staphylococcal bioilm on ISSN
CaP bone
grafts
European
M ClaussCells
et al.and Materials Vol. 28 2014 (pages 39-50)
INFLUENCE OF PHYSICO-CHEMICAL MATERIAL CHARACTERISTICS
ON STAPHYLOCOCCAL BIOFILM FORMATION – A QUALITATIVE AND
QUANTITATIVE IN VITRO ANALYSIS OF FIVE DIFFERENT CALCIUM PHOSPHATE
BONE GRAFTS
M. Clauss1,2,3,*, U. Furustrand Tain4, B. Betrisey3, N. van Garderen2, A. Trampuz3,5, T. Ilchmann1 and M. Bohner2
Clinic for Orthopaedics and Trauma Surgery, Kantonsspital Baselland, Liestal, Switzerland
2
RMS Foundation, Bettlach, Switzerland
3
Infectious Diseases Service, Department of Internal Medicine, University Hospital Lausanne (CHUV),
Lausanne, Switzerland
4
Septic Surgical Unit, University Hospital Lausanne (CHUV), Lausanne Switzerland
5
Department for Traumatology and Reconstructive Surgery, Charité, Berlin, Germany
1
Abstract
Introduction
Various compositions of synthetic calcium phosphates (CaP)
have been proposed and their use has considerably increased
over the past decades. Besides differences in physicochemical properties, resorption and osseointegration,
artiicial CaP bone graft might differ in their resistance against
bioilm formation. We investigated standardised cylinders of
5 different CaP bone grafts (cyclOS, chronOS (both β-TCP
(tricalcium phosphate)), dicalcium phosphate (DCP),
calcium-deicient hydroxyapatite (CDHA) and α-TCP).
Various physico-chemical characterisations e.g., geometrical
density, porosity, and speciic surface area were investigated.
Bioilm formation was carried out in tryptic soy broth (TSB)
and human serum (SE) using Staphylococcus aureus (ATCC
29213) and S. epidermidis RP62A (ATCC 35984). The
amount of bioilm was analysed by an established protocol
using sonication and microcalorimetry. Physico-chemical
characterisation showed marked differences concerning
macro- and micropore size, speciic surface area and porosity
accessible to bacteria between the 5 scaffolds. Bioilm
formation was found on all scaffolds and was comparable
for α-TCP, chronOS, CDHA and DCP at corresponding time
points when the scaffolds were incubated with the same germ
and/or growth media, but much lower for cyclOS. This is
peculiar because cyclOS had an intermediate porosity, mean
pore size, speciic surface area, and porosity accessible to
bacteria. Our results suggest that bioilm formation is not
inluenced by a single physico-chemical parameter alone
but is a multi-step process inluenced by several factors in
parallel. Transfer from in vitro data to clinical situations is
dificult; thus, advocating the use of cyclOS scaffolds over
the four other CaP bone grafts in clinical situations with a
high risk of infection cannot be clearly supported based on
our data.
More than one million patients per year need a bone
grafting procedure to repair a bone defect resulting from a
trauma or a bone disease. The use of autologous cancellous
bone grafts transplanted as fresh bone grafts is regarded
as the gold standard (Delloye et al., 2007; Ketonis et al.,
2010). However, several bone graft substitutes have been
proposed, such as fresh-frozen allogeneic cancellous
bone grafts (Van de Pol et al., 2007; Kappe et al., 2010)
and processed human or bovine cancellous bone grafts
(Tadic and Epple, 2004). All these genuine bone grafts
have a comparable calcium phosphate (CaP) architecture
(Clauss et al., 2013). In the 1970s, various compositions
of synthetic CaP, such as β-tricalcium phosphate (β-TCP)
or hydroxyapatite (HA), were proposed. Their importance
and use have considerably increased over the past decades
(Bohner, 2000). Their subtle differences in composition
and structure may have a profound effect on their in vivo
behaviour (Bohner, 2000). Besides differences in chemical
properties, resorption and osseointegration, artiicial bone
graft might differ in case of bacterial colonisation and
bioilm formation.
Surgical site infection is a recognised and often
devastating complication in orthopaedic surgery ranging
from 0.7-4.2 % in elective orthopaedic surgery (Crockarell
et al., 1998) to 30 % following third degree open fractures
(Ostermann et al., 1994). S. aureus and S. epidermidis
are the main microorganisms responsible for 60-80 % of
these infections (Gristina, 1987; Trampuz and Zimmerli,
2006b). Despite antimicrobial prophylaxis in modern
operating rooms, surgical site infections cannot be
completely prevented (Trampuz and Zimmerli, 2006a),
especially in the vicinity of a foreign body (Busscher et
al., 2012). Bacteria are growing attached to the surface as
bioilm (Costerton et al., 2005; Trampuz and Zimmerli,
2006b). The treatment and eradication of infections caused
by bioilms is more dificult than of bacteria growing in
free-living (planktonic) form (Busscher et al., 2012). The
eradication of infection is often only possible by removal
of the foreign body and long-term antimicrobial treatment
(Ehrlich et al., 2005).
The “race to the surface” (Gristina, 1987) as a
multistep process of initial bacterial adhesion and later
bioilm formation is well established for metal implants
(Gristina, 1987; Oga et al., 1988; Cordero et al., 1994;
Vogely et al., 2000; Harris and Richards, 2006; Schlegel
and Perren, 2006; Harris et al., 2007). In contrast, there
Keywords: Bioilm; calcium phosphate; β-TCP; S. aureus
ATCC 29213; S. epidermidis RP62A ATCC 35984;
microcalorimetry; sonication; bone graft.
*Address for correspondence:
Martin Clauss
Kantonsspital Baselland Liestal,
Clinic for Orthopaedics and Trauma Surgery
Rheinstrasse 26
CH-4410 Liestal, Switzerland
Telephone Number: +41 61 925 3722
FAX Number: +41 61 925 2808
Email:
[email protected]
39
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
are only limited data on bioilm formation on the surface
of different CaP bone graft substitutes, mainly HA and
tricalcium phosphate (Van Blitterswijk et al., 1986a; Van
Blitterswijk et al., 1986b; Van Blitterswijk et al., 1986c;
Jakubowski et al., 2008; Westas et al., 2014). Considering
their increasing use in orthopaedic surgery, there might be
more infected bone grafts in the future and differences in
initial adhesion and bioilm formation might have a direct
impact on clinical infection rates. Therefore, the aim of this
study was to analyse the inluence of material properties of
5 different synthetic CaP bone grafts on the initial adhesion
and bioilm formation in an established in vitro setting
(Clauss et al., 2010) using standard laboratory strains of
S. aureus and S. epidermidis which are commonly causing
surgical site bone infection.
device (VELP Scientiica, Usmate, Italy). The samples
were not sintered.
The ifth CaP material (named here “cyclOS”) was
a β-TCP scaffold which was purchased at Mathys Ltd
(Bettlach, Switzerland). According to the producer, these
materials were obtained by pressing a mixture of tricalcium
phosphate powder and polymer beads (125-625 µm),
followed by sintering at 1100 ºC.
Physico-chemical characterisation
Various physico-chemical characterisations were
performed on all the investigated scaffolds: measurement
of their weight, diameter, and height to determine their
geometrical density and porosity, X-ray diffraction (XRD)
to determine their crystalline composition, Nitrogen
adsorption to determine their speciic surface area (SSA)
using the BET model, mercury porosimetry to determine
their microstructure in the range between 7 nm and
100 μm using the Washburn equation, scanning electron
microscopy (SEM) and optical microscopy to look at their
morphology and estimate the macropore size (pore size
> 50-100 µm). More details are provided below.
The geometrical density was determined by dividing
the weight by the apparent volume of the scaffolds. For
XRD the scaffold was crushed, homogenised and packed
in a cavity in an aluminium sample holder. XRD data were
collected in relective geometry on a Panalytical CubiX
diffractometer (Panalytical, Eindhoven, The Netherlands)
equipped with a graphite monochromator in the secondary
beam. CuKα radiation and a step size of 0.02° were
used to measure from 4.01 to 59.99° 2θ. Quantitative
phase analysis was done by Rietveld reinement with the
computer program FullProf.2k (Version 5.00) (RodriguezCarvajal, 2001), using a previously determined instrument
resolution function. Starting models for the quantiied
phases were taken from Dickens et al. (Dickens et al., 1974)
for β-TCP, Mathew et al. (Mathew et al., 1977) for α-TCP,
Sudarsanan and Young (Sudarsanan and Young, 1969) for
hydroxyapatite, Dickens et al. (Dickens et al., 1971) for
DCP, and Curry and Jones (Curry and Jones, 1971) for
dicalcium phosphate dihydrate (DCPD; CaHPO4·2H2O).
The SSA was measured by nitrogen adsorption using the
BET model (Gemini 2360, Micromeritics, Norcross, GA,
USA). Four measurements were made per material. Total
porosity and pore size distributions were evaluated with
a mercury porosimeter up to 200 MPa (Pascal 140/440,
Thermo Fisher, Schwerte, Germany). Surface tension and
contact angle of mercury were set to 0.480 N/m and 140°,
respectively. Samples were dried overnight at 130 °C in
order to drive off any physisorbed water from the sample.
Three different samples were measured to determine the
standard deviation. For SEM, broken pieces of the scaffolds
were placed on a sticky carbon tape, itself sticking on an
aluminium sample holder. The particles were then sputtered
with C and subsequently with Au to a total thickness of
approximately 20 nm. The samples were observed with an
EVO MA25 microscope (Zeiss, Oberkochen, Germany).
The macropore size of the samples was estimated by optical
microscopy using a method previously described (Bohner
et al., 2001). Briely, photos of the surface of polished and
cleaned scaffolds were taken (Leica (Wetzlar, Germany).
Material and Methods
Four out of 5 examined CaP scaffolds used in the present
study were produced using the so-called calcium phosphate
emulsion method (Bohner, 2001; Kasten et al., 2003;
Bohner et al., 2005). In brief, a powder mixture consisting
of 80 g of α-tricalcium phosphate (α-TCP; α-Ca3(PO4)2;
produced in-house) and 20 g of tricalcium phosphate
(Merck, Dietikon, Switzerland) was mixed with 100 g
viscous parafin oil (Merck) and 60 mL of 0.2 M Na2HPO4
aqueous solution containing 0.57 g/L of polyethoxylated
castor oil (Cremophor EL, BASF, Wädenswil, Switzerland),
and 1 % of 5.1 kDa poly(acrylic acid) (Fluka, Buchs,
Switzerland). After 2 min of mixing at 2000 rpm, the paste
was poured into standing 30 mL syringes the tip of which
had been pre-cut. Fourty-ive minutes later, the samples
were covered with 10 mL phosphate buffer (phosphatebuffered saline, PBS) (4.5 g/L NaCl, 1.79 g/L KH2PO4,
9.0 g/L Na2HPO4·2H2O), and incubated for one day at
60 ºC. Then the samples were incubated in petroleum ether,
dried and sintered at 1250 ºC. After drying, the samples
were lathed to obtain the desired block dimensions. The
last processing steps included a cleaning stage in ethanol
and inal calcination at 900 ºC for 1 h.
The samples obtained according to this procedure
consisted of pure β-tricalcium phosphate (β-TCP;
β-Ca3(PO4)2, named here “chronOS” due to their similarity
to the product sold by DePuy Synthes (West Chester, PA,
USA). To obtain α-TCP scaffolds, the β-TCP scaffolds
(chronOS) were calcined at 1500 ºC for 12 h and rapidly
cooled down to room temperature. The conversion of
α-TCP scaffolds into dicalcium phosphate scaffolds (DCP;
CaHPO4) occurred via a chemical reaction between α-TCP
scaffolds and a phosphoric acid solution (Galea et al.,
2008).
For the synthesis of calcium-deicient hydroxyapatite
(CDHA; Ca 9(PO 4) 5(HPO 4)(OH)), calcium phosphate
emulsions were also used (Kasten et al., 2003). Compared
to the synthesis of β-TCP scaffolds, polyacrylic acid was
replaced by sodium citrate (0.2 M concentration) and the
fraction of emulsiier was reduced from 0.57 to 0.40 g/L.
The incubation at 60 ºC lasted 72 h instead of 24 h and the
samples were cleaned with petroleum ether in a soxhlet
40
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
sample to another. Afterwards samples were split in two
groups for further analysis: Group A was analysed by
microcalorimetry alone and group B by sonication and
microcalorimetry.
(ii) Sonication procedure (only group B): After
washing, samples were transferred to new 50 mL-Falcon
tubes containing 5 mL 1 % PBS, gently shaken for 5 s,
sonicated at 40 kHz for 5 min in a bath tub sonicator
(BactoSonic, Bandelin, Germany) and shaken again for 5 s.
The dislodged bioilm (sonication luid) was transferred
to a 10 mL-Falcon tube and CaP bone grafts were stored
for microcalorimetry (see below).
Sonication luid was serially diluted in Eppendorf tubes
and aliquots of 100 µL were plated on sheep blood agar and
incubated at 37 °C aerobically for 24 h. Bacterial counts
were enumerated and expressed as cfu/sample. Plates were
examined for variations in colony morphology (colour,
size) and contaminations.
(iii) Microcalorimetry protocol (group A and B): all
microcalorimetry tests were performed using a 48-channel
batch calorimeter (thermal activity monitor, model 3102
TAM III; TA Instruments, New Castle, DE, USA).
In more detail, biofilm-loaded CaP samples were
transferred into sterile 4 mL calorimeter ampoules prefilled with 1 mL TSB, closed with a rubber cap and
sealed by manual crimping. Ampoules were sequentially
introduced into the microcalorimeter and remained for
15 min in the thermal equilibration position before they
were lowered into the measurement position. Heat low
was measured continuously after the signal stability was
achieved throughout an 18 h period and expressed as heat
low over time (in microwatts [µW]). The calorimetric time
to detection (TTD) was deined as the time from insertion
of the ampoule into the calorimeter until the exponentially
rising heat low signal exceeded 50 µW to distinguish
microbial heat production from the thermal background.
TTD indirectly quantiies the amount of bacteria with
a shorter TTD representing a larger amount of bacteria.
Data analysis was accomplished using the manufacturer’s
software (TAM Assistant; TA Instruments) and Prism 5.0
(GraphPad Software, La Jolla, CA).
MZ12 microscope, JVC KY-F70 digital camera,
Image Access software). The average diameter of ifteen
macropores was determined and an average macropore
diameter, De, was calculated. The inal macropore diameter,
D, was calculated from De using equation (1) (assuming
that all the macropores are round and homogeneously
distributed):
(1)
Bioilm formation
Two established bioilm forming staphylococcal strains
were used (Clauss et al., 2010): S. aureus (ATCC 29213)
is a gram-positive, coagulase-positive, methicillinsusceptible, bioilm-forming strain (Ceri et al., 2001). S.
epidermidis RP62A (ATCC 35984) is a gram-positive,
coagulase-negative, bioilm-forming strain (Merritt et
al., 1998). The strains were stored at -70 °C by using
a cryovial bead preservation system (Microbank, Por.
Lab Diagnostics, Richmond Hill, Ontario, Canada). For
preparation of the inoculum, a single bead was freshly
grown on sheep blood agar overnight. Bacterial inocula
were prepared from discrete colonies resuspended in 1 %
PBS to a McFarland turbidity of 0.5 representing a bacterial
concentration of ~1.0 x 108 colony-forming units (cfu)/
mL. The stock solution was diluted 1:1.000 for further
experiments.
Bioilm formation was performed in (i) tryptic soy
broth (TSB, Beton Dickinson AG, Basel Switzerland) and
(ii) undiluted pooled human serum (serum, Millipore™
Temecula CA, United States) as recently published (Clauss
et al., 2010). In brief, CaP bone grafts were inserted in
50 mL-Falcon tubes pre-illed with 2700 µL of medium.
To allow a homogeneous soaking of the porous materials
over a 30 min period, samples were placed on top of the
liquid surface (Stähli et al., 2010). At the end of the period,
samples were completely submerged due to the additional
water content in the pores (Clauss et al., 2013). In a inal
step, 300 µL of diluted bacterial stock solution were added
resulting in an initial bacterial concentration of S. aureus
of ~1.5 x 105 cfu/mL, and for S. epidermidis ~1.0 x 105
cfu/mL. Samples were incubated under static conditions
at 37 °C with ambient air for either 3 h, 24 h or 72 h.
Biofilm analysis was adapted from our recently
published procedure (Clauss et al., 2010). Three steps were
included: (i) washing procedure (all samples) followed
by (ii) sonication (half of the samples) and (iii) a inal
microcalorimetric analysis (all samples).
(i) Washing procedure: After incubation, CaP scaffolds
were transferred to a new 50 mL-Falcon tube (preilled
with 5 mL 1 % PBS) with a sterile forceps. They were
washed 5 times with 5 mL 1 % PBS to remove planktonic
bacteria. For washing the PBS was poured in the Falcon
tubes by placing a glass pipette on the wall of the Falcon
tubes; afterwards the Falcon tubes were shaken cautiously
by hand, and in a inal step the PBS was sucked-off by
placing a Pasteur pipette on top, to one side of the CaP
scaffolds, to have a lush through the scaffold. Both the
glass pipette and the Pasteur pipette were changed after
processing one scaffold, to avoid contamination from one
Statistical calculations
Physico-chemical characterisation was done with at least
three scaffolds of each material. Data are presented as mean
and standard deviation (SD).
Bioilm experiments were performed in triplicates. To
equalise variances in bacterial counts, data are presented
as log10 cfu/sample. For statistical analysis a one-way
ANOVA with Bonferroni’s multiple comparison test was
performed using Prism 5.0 (GraphPad Software, La Jolla,
CA). A p-value of < 0.05 was considered to be signiicant.
Results
Material characteristics
Apart from the DCP samples, all scaffolds had a purity
close to 100 % (Table 1). The DCP scaffolds contained
about 6 % of remaining or unreacted α-TCP. All samples
contained large pores (macropores) with a diameter close
41
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
Fig. 1. SEM images of the 5 investigated CaP scaffold types. Three enlargements were used per material, from a
small enlargement (images on top; scale bar: 200 µm) to a more detailed analysis of broken surfaces (images at the
bottom; scale bar: 10 µm). The intermediate enlargements (middle images) have a scale bar of 50 µm.
Table 1. Summary of the physico-chemical properties of the samples used in the present study. Crystalline composition
(Rietveld reinement analysis of the XRD data), speciic surface area (SSA), macropore diameter, apparent density,
porosity, median pore size (d50) and porosity accessible by bacteria (> 1.5 μm) in mean and standard deviation.
Speciic
surface
area
2
Material
Composition
[m /g]
β-TCP (chronOS)
>99 %1)
0.32 ± 0.01
β-TCP (cyclOS)
>99 %2)
0.84 ± 0.15
α-TCP
>99 %
0.16 ± 0.103)
93 % DCP,
DCP
6 % α-TCP, 4.04 ± 0.35
1 % DCPD
98 % HA,
CDHA
43.6 ± 0.4
2 % DCP
Porosity
accessible
by
Macropore
bacteria
Diameter App Density Porosity Porosity* (> 1.5 μm) *
d50 *
[mm]
0.39 ± 0.09
0.26 ± 0.07
0.41 ± 0.09
[g/cm ]
0.84 ± 0.03
0.88 ± 0.03
0.89 ± 0.02
[%]
72.6 ± 1.0
71.1 ± 1.0
68.9 ± 0.7
[%]
68 ± 3
70 ± 3
62 ± 1
[%]
59 ± 3
59 ± 3
61 ± 1
[μm]
16 ± 3
17 ± 3
51 ± 4
0.37 ± 0.08
1.17 ± 0.04
60.0 ± 1.4
46 ± 2
37 ± 2
27± 7
0.53 ± 0.13
0.53 ± 0.01
82.0 ± 0.3
69 ± 4
27 ± 9
0.23 ±
0.18
3
* Determined by mercury porosimetry
1)
Crystallite size: 163 ± 34 nm (± 1 St Dev)
2)
Crystallite size: 103 ± 12 nm (± 1 St Dev)
3)
This value is at the lower range of what the instrument can measure
Fig. 2. Representative curves showing
the normalised pore size distributions of
the 5 investigated CaP scaffold types. The
curves represent the fraction of the pores
smaller than a given value. The vertical
line corresponds to the approximate size
of bacteria.
42
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
a)
b)
Fig. 3. The cfu count in sonication luid of scaffolds incubated in (a) TSB and (b) serum (SA S. aureus, SE S.
epidermidis). Data presented as mean ± SD.
Contrarily, large differences in micropore size were
observed by SEM. Qualitatively, the mean pore size as
determined by Hg porosimetry decreased in the order:
α-TCP > chronOS > DCP > cyclOS > CDHA (Table 1).
Microstructure in terms of pore size distribution was
further studied by Hg porosimetry to evaluate the possible
bacteria invasion into the structure (Fig 2). Pore size
distributions were normalised according to their volume in
order to show the differences between scaffolds. According
to these results, all materials but α-TCP had pores small
enough to partly prevent bacterial invasion (estimated size
of the two bacterial strains: 1.5 μm). Whereas the porosity
fraction, not accessible to bacteria, was close to 80-87 %
for chronOS, cyclOS and DCP, this value dropped to 39 %
for CDHA scaffolds (= 27 %/69 %).
The porosities determined by Hg porosimetry were 1 to
14 % smaller than the values obtained via the determination
of the geometrical density (Table 1). Whereas the
differences were small for chronOS and cyclOS (4 % and
1 %), much larger differences were observed for DCP
(14 %) and CDHA (13 %).
to 0.2-0.6 mm and smaller pores, either in the nm (e.g.,
for CHDA and DCP) or in the low μm range (1-10 μm)
(Table 1, Fig. 1).
α-TCP and DCP scaffolds were produced from β-TCP
scaffolds (chronOS), so their macropore size distribution
and morphology were similar (Fig. 1; irst row). However,
the macropores of DCP samples were partly illed with
small DCP protuberances, which led to a 10 % lower
porosity and 5-10 % decrease in macropore size compared
to β-TCP and α-TCP scaffolds (Table 1). The difference
of microstructure seen on the SEM photos between the
various scaffolds was relected by large variations of SSA
values. Indeed, the SSA values of the scaffolds varied over
a very wide range, starting from 0.16 ± 0.10 m2/g for α-TCP
scaffolds up to 43.6 ± 0.4 m2/g for CDHA scaffolds.
SEM images also revealed that CDHA samples had the
largest macropore size (0.53 ± 0.13 mm), whereas cyclOS
samples had the smallest macropore size (0.26 ± 0.07 mm).
These observations were supported by the macropore
size estimated from polished scaffold surfaces (Table 1).
However, changes in macropore size remained small.
43
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
Fig. 4. microcalorimetric results from
chronOS scaffolds incubated for 3 h
with S. aureus and TSB (black line), S.
aureus and serum (black dotted line), S.
epidermidis and TSB (red line) and S.
Bioilm formation
Both test strains formed a bioilm on all test materials,
as conirmed by quantitative culture after washing and
sonication of the CaP scaffolds. The amount of bioilm
removed from the scaffolds showed low variation within
a triplicate and was comparable for 4 of the 5 tested
materials, namely α-TCP, chronOS, CDHA and DCP but
different for cyclOS (Fig. 3a,b).
For α-TCP, chronOS, CDHA and DCP incubated
with S. aureus in TSB, differences between 3 h and 24 h
(p < 0.001), 3 h and 24 h (p < 0.001) and 24 h and 72 h
(p > 0.05) were comparable. The differences were the same
when these 4 materials were incubated with S. aureus in
serum, and when incubated with S. epidermidis in TSB
(Fig. 3).
For cyclOS incubated with S. aureus in TSB, differences
were less marked between 3 h and 24 h (p < 0.01), 3 h
and 24 h (p < 0.05) and 24 h and 72 h (p > 0.05). When
cyclOS was incubated in serum there were no signiicant
differences (p > 0.05) between the 3 time points (Fig. 3a).
When the 5 materials were incubated with S.
epidermidis in serum, again no signiicant differences
(p > 0.05) were found between the 3 time points for cyclOS,
but a continuous and highly signiicant increase was noticed
between 3 h and 24 h for α-TCP (p < 0.001), chronOS
(p < 0.01), CDHA (p < 0.001), and DCP (p < 0.001), and
between 24 h and 72 h for α-TCP (p < 0.001), chronOS
(p < 0.001), and DCP (p < 0.001) (Fig. 3b).
The qualitative analysis of the biofilm showed a
homogenous size of CFU in the sonication luid for all
samples without contaminations on the plates.
The triplicates in microcalorimetric measurements
showed a high uniformity of the shape of the curves
(indicating no gross contamination on the scaffolds) with
low variations in TTD, independent from the strain, growth
medium, and length of incubation (Fig. 4). When these
indings are combined with the qualitative analysis of the
sonication luid, contaminations on the samples can be
excluded.
The lowest number of bacteria (longest TTD) was found
on the surface of the CDHA scaffolds (p < 0.001 S. aureus
and S. epidermidis) after 3 h incubation in TSB (Fig. 5a).
The differences decreased after 24 h incubation (p > 0.05
S.aureus, p < 0.05 S. epidermidis) and disappeared after
72 h incubation (p > 0.05). When incubated in serum,
there was signiicantly less bioilm on the surface of the
CDHA and the cyclOS scaffolds as compared to the three
other samples after 3 h incubation (p < 0.001). After 24 h
and 72 h incubation cyclOS showed the lowest amount of
bioilm (Fig. 5).
When the samples from group A (washing) and group
B (washing and sonication) were compared less bioilm
(longer TTD) was detected on the surface of all samples,
but no change in the relative differences between the
materials, time-points, and media (Fig. 5). All materials
showed less bioilm (longer TTD) at corresponding time
points when incubated in serum as compared to TSB (Fig.
5).
The amount of bioilm detected by microcalorimetry
(TTD) for group A (washing) and B (washing and
sonication) correlated well with the cfu counts after
sonication.
Discussion
Numerous studies have been devoted to the quest of an ideal
bone scaffold design (pore size, interconnectivity, shape,
stability) without giving a clear answer (Bohner et al.,
2011). Furthermore, depending on the clinical application,
these speciications might vary. Design modiications may
not only change bony ingrowth and substitution with time
(Bohner et al., 2011) but also inluence bacterial adhesion
and the related risk of infection (Clauss et al., 2013)
because they use the same adhesive mechanisms (Busscher
et al., 2012). From a clinical perspective, initial bacterial
colonisation, competing with tissue cell integration, better
known as the “race to the surface” (Gristina, 1987), is the
most important step as it preludes biomaterial associated
infection (Busscher et al., 2012).
Extensive research has been performed to determine the
propensity of medical devices to sustain bioilm formation
by staphylococci (Busscher et al., 2012; Clauss et al.,
2010; Harris and Richards, 2006). It has been shown, that
various physico-chemical factors inluence in vitro bioilm
formation on a surface (Harris et al., 2007; Jakubowski
44
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
a)
b)
Fig. 5. Results from microcalorimetry. (a) washing procedure (group A), (b) washing and sonication (group B). Data
presented as mean ± SD.
clinically were considered. These two materials present
very similar pore architectures. In both cases, the pores are
spherical, the porosity is close to 70 %, and the structure
is highly microporous (Figs. 1 and 2). The speciic surface
area is larger, whereas the macropores are slightly smaller
and less interconnected for cyclOS compared to chronOS.
CDHA is generally not used as scaffold due to its more
complex manufacturing process (Steffen et al., 2001;
Kasten et al., 2003), but since it is the end product of the
setting reaction of many calcium phosphate cements, it
is a relevant material. Also, CDHA is a very interesting
material to analyse considering the importance of surfaces
for bacterial bioilms. Indeed, its speciic surface area is
several orders of magnitude larger than that of β-TCP and
α-TCP scaffolds (Table 1). Whereas CDHA and β-TCP
et al., 2008; MacKintosh et al., 2006; Oga et al., 1988;
Patel et al., 2003; Patel et al., 2007). There is only limited
information about staphylococcal bioilm formation on the
surface of CaP bone grafts investigating S. aureus bioilm
formation on either HA (Van Blitterswijk et al., 1986a; Van
Blitterswijk et al., 1986b) or β-TCP (Van Blitterswijk et
al., 1986c) in the middle ear of rats.
The aim of this study was to analyse the inluence
of physico-chemical and structural variations of CaP
bone grafts on staphylococcal bioilm formation under
standardised in vitro conditions.
Several CaP materials were selected based on their
clinical and scientiic relevance. Nowadays, β-TCP is
perhaps the most commonly used CaP bone graft substitute
(Bohner, 2010). Therefore, two β-TCP scaffolds used
45
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
the difference in growth between the two culture media. An
interesting observation was that less bioilm was found on
the cyclOS scaffolds both when TSB and serum were used,
which proved that the smaller amount of bioilm found on
this particular material was not due to interaction between
the growth media (TSB or serum) and the material.
Various methods for quantitative/qualitative evaluation
of bioilm formation, like “live-dead staining”, confocal
laser scanning microscopy (CLSM) or SEM have been
described. All methods except calorimetry need a special
pre-treatment, such as staining (live-dead staining, CLSM)
or carbon-sputtering (SEM), which interfere with further
bioilm investigation (Clauss et al., 2010). The surface
that can be monitored by other methods is restricted to
small areas while microcalorimetry monitors the whole
device (Clauss et al., 2010). For microcalorimetric
monitoring of bioilm formation it is essential to remove
planktonic bacteria from the scaffolds, and washing might
be critical with porous materials such as CaP bone grafts.
Planktonic bacteria might remain in the pores, resulting
in an overestimation of the true amount of bioilm. To
minimise this effect we modiied our recently published
washing protocol (Clauss et al., 2010) for removal of
planktonic bacteria. Uniformly, all samples were washed
ive times placing the suction device directly on top, and
to one side, of the sample producing a lush through the
scaffolds. We cannot quantify the amount of remaining
planktonic bacteria but variations in cfu counts and TTD
within a triplicate have been low, thus systematic error
should be low and comparable on all scaffolds.
Several material parameters are known to modify the
rate of infection in vitro. For example, porous materials
have a higher rate of infection than dense materials
(Merritt et al., 1998; Harris et al., 2007; Clauss et al.,
2010). Rough materials are also more prone to infection
(Lange et al., 2002; Meredith et al., 2005). The investigated
CaP bone grafts were porous, with a microscopically
and macroscopically rough surface. The materials
presented obvious but moderate differences in porosity
and roughness, as revealed by SEM, Hg porosimetry, and
nitrogen adsorption. Bioilm formation on the surface of the
CaP samples by means of sonication and microcalorimetry
was comparable for α-TCP, chronOS, CDHA and DCP,
but much lower for cyclOS. This is peculiar because this
calcium phosphate had an intermediate porosity (Table
1), mean pore size, speciic surface area, and porosity
accessible to bacteria. Therefore, it is unclear why cyclOS
would be less prone to bioilm formation than the other
materials.
Hydrophobicity has been shown to be an important
factor for initial bacterial adhesion and E. coli bioilm
formation on the surface of HA and β-TCP (Jakubowski
et al., 2008).
In the past, Staehli et al. had shown that cyclOS samples
were more dificult to impregnate by a luid than chronOS
(Stähli et al., 2010), what might result in a higher volume
of air in the cyclOS samples reducing the surface available
for bioilm formation. They established an impregnation
test setup and assessed the effect of various synthesis
parameters, such as sintering temperature, composition,
macroporosity and macropore size on the impregnation
solubilities are very similar, the two last materials, i.e., DCP
and α-TCP, have a much higher solubility (Chow, 1991),
which may affect their bacterial reaction. Furthermore,
DCP is considered to be one of the most promising bone
graft substitute materials (Habibovic et al., 2008; Bohner,
2010).
All scaffolds are composed of pores large enough and
interconnected enough to allow the invasion of bacteria
(Ø 0.5-1.5 µm) into the structure. However, differences
can be found between scaffolds: whereas only 39 % of
CDHA porosity can be invaded by 1.5 μm spheres, this
value increases to 80-87 % for DCP, chronOS and cyclOS,
and reaches 100 % for α-TCP.
The two scaffold types with the lowest porosity
accessible by bacteria (CDHA and DCP) are also the
scaffolds with the highest SSA values and presenting
the largest differences between the porosity inferred
from the geometrical density and from Hg porosimetry
measurements. These results suggest that CDHA and DCP
scaffolds are partly crushed during Hg impregnation, hence
leading to an overestimation of the total porosity, and an
error in the porosity fraction smaller than 1.5 μm. Indeed,
the scaffolds used in the present study have a compressive
strength inferior to 10 MPa (particularly CDHA), which
is the pressure required to invade 0.15 μm pores. At this
pressure (or this pore size), only ≈50 % of the CDHA pores
are invaded by Hg (and ≈90 % of the DCP pores).
Choosing the appropriate strain is one of the most
crucial steps when biofilm formation is investigated.
Clinical isolates are not well characterised and results
might be dificult to compare to the literature. We therefore
used standard (ATCC) strains which are known to produce
reproducible amounts of bioilm already after 3 h for S.
aureus (Ceri et al., 1999) and S. epidermidis (Polonio et al.,
2001; Chaw et al., 2005; Qin et al., 2007), at the surface
of various bone grafts (Clauss et al., 2010; Clauss et al.,
2013).
Two different growth media (TSB and human serum)
were chosen because it has been shown that growth media
have a considerable inluence on bioilm formation, as
they change the initial bacterial adhesion on the surface
(Barton et al., 1996; Patel et al., 2007). TSB was frequently
used in previous studies (Clauss et al., 2010; Clauss et
al., 2013), and additionally we used normal undiluted
pooled human serum which was not heat-inactivated by
the manufacturer. The presence of bacterial bioilms grown
in serum proves that the two staphylococcal strains tested
were able to survive and proliferate (more bacteria after
24 h and 72 h compared to 3 h of incubation). Compared to
the dislodged bioilms that had been cultured in TSB (Fig.
3a), the bioilms grown in serum were slightly smaller (i.e.,
lower cfu counts obtained). This could be explained by the
fact that the TSB is a very nutrient-rich growth medium in
comparison to serum. We believe that in vitro experiments
performed in serum better represent the clinical setting,
where the bacteria are present in a more challenging and
harsh environment, compared to the nutrient-rich TSB.
The goal of this study was to investigate the effect
of different physico-chemical material characteristics on
staphylococcal bioilm formation, and not to compare
different growth media. Therefore, we did not emphasise
46
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
properties of porous CaP scaffolds dipped in water.
Among those parameters, the macropore size had by far
the largest effect; generally, the bigger the macropore size,
the lower the saturation level. The results also showed that
impregnation was less complete when the samples were
fully dipped in water than when they were only partially
dipped, owing to the requirement for the system to create
air bubbles under water. We were not able to modify pore
size of the samples as this was one of our major issues
in the present study, but we placed the scaffolds on top
of the surface of the liquid allowing them to submerge
within 30 min, wetting just as much surface as possible.
One would also expect that the effect of entrapped air
should reduce with time, thus the inluence on the results
should become smaller from the 3 h to 24 h and 72 h
incubation. As differences in bioilm formation became
more obvious by longer incubation intervals, differences
in impregnation can most likely not explain the observed
differences in bioilm formation. Besides hydrophobicity,
bacterial adhesion can also be inluenced by the charge of
the surface, i.e., negatively charged bacteria adhere better
to a positively charged surface (Li and Logan, 2004). We
did not investigate surface charge, thus we cannot exclude
that surface charge inluenced our results.
SEM studies and L. Galea for the BET and XRD analysis.
Furthermore we have to thank S. Gersbach (Kantonsspital
Baselland Liestal) for her help with data management and
analysis.
References
Barton AJ, Sagers RD, Pitt WG (1996) Bacterial
adhesion to orthopedic implant polymers. J Biomed Mater
Res 30: 403-410.
Bohner M (2000) Calcium orthophosphates in
medicine: from ceramics to calcium phosphate cements.
Injury 31 Suppl 4: 37-47.
Bohner M (2001) Calcium phosphate emulsions:
possible applications. Key Eng Mater 192-195: 765-768.
Bohner M (2010) Resorbable biomaterials as bone graft
substitutes. Mat Today 13: 24-30.
Bohner M, Van Lenthe GH, Grünenfelder S, Hirsiger W,
Evison R, Müller R (2005) Synthesis and characterization
of porous beta-tricalcium phosphate blocks. Biomaterials
26: 6099.
Bohner M, Loosli Y, Baroud G, Lacroix D (2011)
Commentary: Deciphering the link between architecture
and biological response of a bone graft substitute. Acta
Biomater 7: 478-484.
Busscher HJ, van der Mei HC, Subbiahdoss G, Jutte
PC, van den Dungen JJ, Zaat SA, Schultz MJ, Grainger
DW (2012) Biomaterial-associated infection: locating the
inish line in the race for the surface. Sci Transl Med 4:
153rv110.
Ceri H, Olson M, Morck D, Storey D, Read R, Buret
A, Olson B (2001)The MBEC assay system: Multiple
equivalent bioilms for antibiotic and biocide susceptibility
testing. In: Methods Enzymol 337: 377-385.
Ceri H, Olson ME, Stremick C, Read RR, Morck D, Buret
A (1999) The Calgary Bioilm Device: New technology
for rapid determination of antibiotic susceptibilities of
bacterial bioilms. J Clin Microbiol 37: 1771-1776.
Chaw KC, Manimaran M, Tay FEH (2005) Role
of silver ions in destabilization of intermolecular
adhesion forces measured by atomic force microscopy in
Staphylococcus epidermidis bioilms. Antimicrob Agents
Chemother 49: 4853-4859.
Chow LC (1991) Development of self-setting calcium
phosphate cements. Japan Ceram Soc 99: 954-964.
Clauss M, Trampuz A, Borens O, Bohner M, Ilchmann
T (2010) Bioilm formation on bone grafts and bone graft
substitutes: comparison of different materials by a standard
in vitro test and microcalorimetry. Acta Biomater 6: 37913797.
Clauss M, Tain UF, Bizzini A, Trampuz A, Ilchmann
T (2013) Bioilm formation by staphylococci on fresh,
fresh-frozen and processed human and bovine bone grafts.
Eur Cell Mater 25: 159-166.
Cordero J, Munuera L, Folgueira MD (1994) Inluence
of metal implants on infection. An experimental study in
rabbits. J Bone Joint Surg Br 76: 717-720.
Costerton JW, Montanaro L, Arciola CR (2005) Bioilm
in implant infections: its production and regulation. Int J
Artif Organs 28: 1062-1068.
Conclusion
All investigated CaP bone grafts showed significant
amounts of bioilm on the surface already after 24 h. CyclOS
had the lowest bioilm formation, but this difference cannot
be explained rationally based on the physico-chemical
characterisations such as pore size distribution, porosity,
speciic surface area. Our results suggest that bioilm
formation is not inluenced by a single physico-chemical
parameter alone, but is a multi-step process inluenced
by several factors in parallel. Further studies are needed
to decipher the whole cascade. Infections on the surface
of medical devices are mainly acquired perioperative,
thus a high resistance against initial bacterial adhesion
and bioilm formation is favourable to reducing infection
rates with medical devices. Transfer from in vitro data to
the clinical situation is dificult. As there might be various
reasons for reduced bioilm formation on cyclOS scaffolds,
compared with the four other CaP bone grafts, we believe
that advocating cyclOs in clinical situations with a high
risk of infection such as e.g., open fractures, cannot clearly
be supported based on our data.
Acknowledgements
This study was supported by research grants from the RMS
Foundation; Bettlach, Switzerland (E09_0001), the Swiss
Society of Orthopaedics and Traumatology (SGOT/SSOT)
and 3R Foundation (S124-10). S. epidermidis RP62A
(ATCC 35984) was kindly provided by Prof. P. Vaudaux
from the Department of Infectious Disease, University
Hospital, Geneva, Switzerland. At the RMS Foundation we
want to thank S. Grünefelder and P. Brotschi for their help
producing the scaffolds, W. Hirsinger for his help with the
47
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
Crockarell JR, Hanssen AD, Osmon DR, Morrey
BF (1998) Treatment of infection with debridement and
retention of the components following hip arthroplasty. J
Bone Joint Surg Am 80: 1306-1313.
Curry NA, Jones DW (1971) Crystal structure of
brushite, calcium hydrogen orthophosphate dihydrate: a
neutron diffraction investigation. J Chem Soc (A) Inorg
Phys Theor 1971: 3725-3729.
Delloye C, Cornu O, Druez V, Barbier O (2007) Bone
allografts: What they can offer and what they cannot. J
Bone Joint Surg Br 89: 574-579.
Dickens B, Bowen JS, Brown BE (1971) A reinement
of the crystal structure of CaHPO4 (synthetic monetite).
Acta Cryst B28: 797-806.
Dickens B, Schroeder LW, Brown BE (1974)
Crystallographic studies of the role of Mg as a stabilizing
impurity in β-Ca3(PO4)2. I. The crystal structure of pure
β-Ca3(PO4)2. J Solid State Chem 10: 232-248.
Ehrlich GD, Stoodley P, Kathju S, Zhao Y, McLeod BR,
Balaban N, Hu FZ, Sotereanos NG, Costerton JW, Stewart
PS, Post JC, Lin Q (2005) Engineering approaches for the
detection and control of orthopaedic bioilm infections.
Clin Orthop Relat Res 437: 59-66.
Galea LG, Bohner M, Lemaitre J, Kohler T, Muller
R (2008) Bone substitute: transforming beta-tricalcium
phosphate porous scaffolds into monetite. Biomaterials
29: 3400-3407.
Gristina AG (1987) Biomaterial-centered infection:
Microbial adhesion versus tissue integration. Science 237:
1588-1595.
Habibovic P, Gbureck U, Doillon CJ, Bassett DC, van
Blitterswijk CA, Barralet JE (2008) Osteoconduction and
osteoinduction of low-temperature 3D printed bioceramic
implants. Biomaterials 29: 944-953.
Harris LG, Richards RG (2006) Staphylococci and
implant surfaces: a review. Injury 37 Suppl 2: S3-14.
Harris LG, Meredith DO, Eschbach L, Richards RG
(2007) Staphylococcus aureus adhesion to standard microrough and electropolished implant materials. J Mater Sci:
Mater Med 18: 1151-1156.
Jakubowski W, lósarczyk A, Paszkiewicz Z, Szymanski
W, Walkowiak B (2008) Bacterial colonisation of
bioceramic surfaces. Adv Appl Ceram 107: 217-222.
Kappe T, Cakir B, Mattes T, Reichel H, Flören M (2010)
Infections after bone allograft surgery: a prospective study
by a hospital bone bank using frozen femoral heads from
living donors. Cell Tissue Banking 11: 253-259.
Kasten P, Luginbuhl R, van Griensven M, Barkhausen
T, Krettek C, Bohner M, Bosch U (2003) Comparison of
human bone marrow stromal cells seeded on calciumdeicient hydroxyapatite, beta-tricalcium phosphate and
demineralized bone matrix. Biomaterials 24: 2593-2603.
Ketonis C, Barr S, Adams CS, Hickok NJ, Parvizi
J (2010) Bacterial colonization of bone allografts:
establishment and effects of antibiotics. Clin Orthop Relat
Res 468: 2113-2121.
Lange R, Luthen F, Beck U, Rychly J, Baumann A,
Nebe B (2002) Cell-extracellular matrix interaction and
physico-chemical characteristics of titanium surfaces
depend on the roughness of the material. Biomol Eng 19:
255-261.
Li B, Logan BE (2004) Bacterial adhesion to glass and
metal-oxide surfaces. Colloids Surf B Biointerfaces 36:
81-90.
MacKintosh EE, Patel JD, Marchant RE, Anderson
JM (2006) Effects of biomaterial surface chemistry on
the adhesion and bioilm formation of Staphylococcus
epidermidis in vitro. J Biomed Mater Res A 78: 836-842.
Mathew M, Schroeder LW, Dickens B, Brown WE
(1977) The crystal structure of a-Ca3(PO4)2. Acta Cryst
B33: 1325-1333.
Meredith DO, Eschbach L, Wood MA, Riehle MO,
Curtis AS, Richards RG (2005) Human ibroblast reactions
to standard and electropolished titanium and Ti-6Al-7Nb,
and electropolished stainless steel. J Biomed Mater Res A
75: 541-555.
Merritt K, Gaind A, Anderson JM (1998) Detection
of bacterial adherence on biomedical polymers. J Biomed
Mater Res 39: 415-422.
Oga M, Sugioka Y, Hobgood CD, Gristina AG,
Myrvik QN (1988) Surgical biomaterials and differential
colonization by Staphylococcus epidermidis. Biomaterials
9: 285-289.
Ostermann PA, Henry SL, Seligson D (1994) Timing of
wound closure in severe compound fractures. Orthopedics
17: 397-399.
Patel JD, Ebert M, Stokes K, Ward R, Anderson JM
(2003) Inhibition of bacterial and leukocyte adhesion under
shear stress conditions by material surface chemistry. J
Biomater Sci Polym Ed 14: 279-295.
Patel JD, Ebert M, Ward R, Anderson JM (2007) S.
epidermidis bioilm formation: Effects of biomaterial
surface chemistry and serum proteins. J Biomed Mater
Res A 80: 742-751.
Polonio RE, Mermel LA, Paquette GE, Sperry JF
(2001) Eradication of bioilm-forming Staphylococcus
epidermidis (RP62A) by a combination of sodium
salicylate and vancomycin. Antimicrob Agents Chemother
45: 3262-3266.
Qin Z, Yang X, Yang L, Jiang J, Ou Y, Molin S, Qu
D (2007) Formation and properties of in vitro bioilms of
ica-negative Staphylococcus epidermidis clinical isolates.
J Med Microbiol 56: 83-93.
Rodriguez-Carvajal J (2001) Recent Developments
of the Program FULLPROF. Commission on Powder
Diffraction (IUCr) Newsletter 26: 12-19.
Schlegel U, Perren SM (2006) Surgical aspects of
infection involving osteosynthesis implants: implant design
and resistance to local infection. Injury 37 Suppl 2: S67-73.
Stähli C, Bohner M, Bashoor-Zadeh M, Doebelin N,
Baroud G (2010) Aqueous impregnation of porous betatricalcium phosphate scaffolds. Acta Biomater 6: 27602772.
Steffen T, Stoll T, Arvinte T, Schenk RK (2001) Porous
tricalcium phosphate and transforming growth factor used
for anterior spine surgery. Eur Spine J 10 Suppl 2: S132140.
Sudarsanan K, Young RA (1969) Signiicant precision
in crystal structure details: Holly springs hydroxyapatite.
Acta Cryst B25: 1534-1543.
Tadic D, Epple M (2004) A thorough physicochemical
characterisation of 14 calcium phosphate-based bone
48
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
substitution materials in comparison to natural bone.
Biomaterials 25: 987-994.
Trampuz A, Zimmerli W (2006a) Antimicrobial agents
in orthopaedic surgery: Prophylaxis and treatment. Drugs
66: 1089-1105.
Trampuz A, Zimmerli W (2006b) Diagnosis and
treatment of infections associated with fracture-ixation
devices. Injury 37 Suppl 2: S59-66.
Van Blitterswijk CA, Bakker D, Grote JJ, Daems ThW
(1986a) The biological performance of calcium phosphate
ceramics in an infected implantation site: II. Biological
evaluation of hydroxyapatite during short-term infection.
J Biomed Mater Res 20: 1003-1015.
Van Blitterswijk CA, Grote JJ, De Groot K (1986b) The
biological performance of calcium phosphate ceramics in
an infected implantation site: I. Biological performance of
hydroxyapatite during Staphylococcus aureus infection. J
Biomed Mater Res 20: 989-1002.
Van Blitterswijk CA, Grote JJ, Koerten HK, Kuijpers W
(1986c) The biological performance of calcium phosphate
ceramics in an infected implantation site III: Biological
performance of β-whitlockite in the noninfected and
infected rat middle ear. J Biomed Mater Res 20: 1197-1217.
Van de Pol GJ, Sturm PDJ, Van Loon CJ, Verhagen C,
Schreurs BW (2007) Microbiological cultures of allografts
of the femoral head just before transplantation. J Bone Joint
Surg B 89: 1225-1228.
Vogely HC, Oosterbos CJ, Puts EW, Nijhof MW,
Nikkels PG, Fleer A, Tonino AJ, Dhert WJ, Verbout AJ
(2000) Effects of hydroxyapatite coating on Ti-6A1-4V
implant-site infection in a rabbit tibial model. J Orthop
Res 18: 485-493.
Westas E, Gillstedt M, Lonn-Stensrud J, Bruzell E,
Andersson M (2014) Bioilm formation on nanostructured
hydroxyapatite-coated titanium. J Biomed Mater Res A
102: 1063-1070.
rates after lesser toe surgery comparing stainless steel and
titanium Kirschner wires and were able to show superior
results (lower recurrence rates, less bioilm formation,
less clinical infections) for the titanium wires (Clauss et
al., 2013). These data conirm the in vitro (Arens et al.,
1996; Harris et al., 2007, additional references) and animal
experimental indings (Harris et al., 2007, text reference;
Melcher et al., 1996; Johansson et al., 1999; Sheehan et
al., 2004; Moriarty et al., 2009, additional references)
published for these two materials. Furthermore we found
some evidence that clinical infection on cerebrospinal luid
catheters is comparable to in vitro and in vivo data (Bayston
and Lambert, 1997; Pattavilakom et al., 2006, additional
references).
However, the question remains “How much importance
should a clinician place on in vitro data or results obtained
from animal models?” In a recent review on S. aureus
osteomyelitis and animal models published, the authors
clearly showed the limitations of various animal models
concerning the transfer from animal data to the clinical
situation (Reizner et al., 2014). We think that choosing a
certain material instead of another will not solely solve the
problem of implant-associated infections alone but might
be one step in reducing the number of infections.
T. Moriarty: In some clinical studies it has emerged that a
signiicant proportion of bacteria that cause device-related
infections are weak bioilm formers in vitro, yet are capable
of causing bioilm infections on indwelling devices in
human patients. Do the authors have any comment on this,
or any data from tests of their materials with weak bioilm
forming strains?
Authors: Bioilm formation is a multi-step process with
some of these steps even running in parallel. The amount
of bioilm formed on the surface thus is not only dependent
to the capability of a bacterial strain to form bioilm or not
but also on the speciic environment. We do not have any
additional in vitro data on bioilm formation with weak
bioilm forming strains on the surface of the investigated
CaP scaffolds.
Discussion with Reviewers
T. Moriarty: Whilst accepting the authors’ statement that
extrapolation of in vitro data to the clinical situation is
dificult, the results presented in this study nevertheless
reveal a measured difference in bacterial bioilm formation
between the test materials. This begs the question “How
much importance should a clinician place on these
results?”. Are the authors able to draw on any examples
of in vitro data of bacterial adhesion or bioilm formation
on any material that has successfully been linked with
clinical infection?
Authors: The reviewer raises two important points: First
concerning the clinical relevance of our presented data.
In the initial submission we draw the conclusion that the
use of cyclOS can be advocated from the in vitro data.
We toned down this statement due to the fact that so far
in vivo data are missing (Geurts et al., 2011, additional
reference). However, it seems reasonable to use cyclOS
instead of another tested CaP scaffold in cases with an
increased risk of infection.
Concerning the second point, we recently published
data on infection rates, bioilm formation and recurrence
Additional References
Arens S, Schlegel U, Printzen G, Ziegler WJ, Perren
SM, Hansis M (1996) Inluence of materials for ixation
implants on local infection. An experimental study of steel
versus titanium DCP in rabbits. J Bone Joint Surg Br 78:
647-651.
Bayston R, Lambert E (1997) Duration of protective
activity of cerebrospinal luid shunt catheters impregnated
with antimicrobial agents to prevent bacterial catheterrelated infection. J Neurosurg 87: 247-251.
Clauss M, Graf S, Gersbach S, Hintermann B, Ilchmann
T, Knupp M (2013) Material and bioilm load of K wires
in toe surgery: titanium versus stainless steel. Clin Orthop
Relat Res 471: 2312-2317.
Geurts J, Chris Arts JJ, Walenkamp GH (2011) Bone
graft substitutes in active or suspected infection. Contraindicated or not? Injury 42 Suppl 2: S82-86.
49
www.ecmjournal.org
Staphylococcal bioilm on CaP bone grafts
M Clauss et al.
Johansson A, Lindgren JU, Nord CE, Svensson O
(1999) Local plate infections in a rabbit model. Injury 30:
587-590.
Melcher GA, Hauke C, Metzdorf A, Perren SM,
Printzen G, Schlegel U, Ziegler WJ (1996) Infection after
intramedullary nailing: an experimental investigation on
rabbits. Injury 27 Suppl 3: SC23-26.
Moriarty TF, Debefve L, Boure L, Campoccia D,
Schlegel U, Richards RG (2009) Inluence of material and
microtopography on the development of local infection
in vivo: experimental investigation in rabbits. Int J Artif
Organs 32: 663-670.
Pattavilakom A, Kotasnas D, Korman TM, Xenos C,
Danks A (2006) Duration of in vivo antimicrobial activity
of antibiotic-impregnated cerebrospinal luid catheters.
Neurosurgery 58: 930-935.
Reizner W, Hunter JG, O’Malley NT, Southgate RD,
Schwarz EM, Kates SL (2014) A systematic review of
animal models for Staphylococcus aureus osteomyelitis.
Eur Cell Mater 27: 196-212.
Sheehan E, McKenna J, Mulhall KJ, Marks P,
McCormack D (2004) Adhesion of Staphylococcus to
orthopaedic metals, an in vivo study. J Orthopaed Res 22:
39-43.
50
www.ecmjournal.org