 HIP

Surgical treatment of Vancouver type B

periprosthetic femoral fractures
PATIENT CHARACTERISTICS AND OUTCOMES OF 1381
­F RACTURES TREATED IN SWEDEN BETWEEN 2001 AND 2011

G. Chatziagorou,
H. Lindahl,
J. Kärrholm
From The Swedish Hip
Arthroplasty Register,
Gothenburg, Sweden
Aims
We investigated patient characteristics and outcomes of Vancouver type B periprosthetic
fractures treated with femoral component revision and/or osteosynthesis.
Patients and Methods
The study utilized data from the Swedish Hip Arthroplasty Register (SHAR) and
information from patient records. We included all primary total hip arthroplasties (THAs)
performed in Sweden since 1979, and undergoing further surgery due to Vancouver type
B periprosthetic femoral fracture between 2001 and 2011. The primary outcome measure
was any further reoperation between 2001 and 2013. Cross-referencing with the National
Patient Register was performed in two stages, in order to identify all surgical procedures
not recorded on the SHAR.

Results
Out of 1381 Vancouver type B fractures that fulfilled the inclusion criteria, 257 underwent
further reoperation by the end of 2013. Interprosthetic and Type B1 fractures had a higher
risk for reoperation. For B1 fractures, the rate of reoperation did not differ (p = 0.322) after
use of conventional (26%) or locking plate osteosynthesis (19%). No significant differences
were observed between cemented, cementless monoblock, and cementless modular
revision components for the treatment of type B2 and B3 fractures.

Conclusion
In this country-specific study, the choice of locking or conventional plates for the treatment
of type B1, and cemented or cementless femoral components fixation for B2 and B3
fractures, had no significant influence on risk for reoperation. Interprosthetic fractures
adversely affected the outcome of treatment of type B fractures. Differences in the patient
characteristics of the compared groups were observed.
Cite this article: Bone Joint J 2019;101-B:1447–1458.

Correspondence should be
sent to G. Chatziagorou; email:
g.chatziagorou@gmail.com
©2019 The British Editorial
Society of Bone & Joint Surgery
doi:10.1302/0301-620X.101B11.
BJJ-2019-0480.R2 $2.00
Bone Joint J
2019;101-B:1447–1458.
THE BONE & JOINT JOURNAL
Fractures of the femoral diaphysis close to a hip
prosthesis are classified as type B, using the Van-
couver classification system.1 They are further
divided into three subcategories: B1, B2, and B3.
In B1 fractures, the femoral component is well
fixed, while in B2 and B3 fractures, the compo-
nent is loose, with (B2) or without (B3) loss of
bone stock. The treatment of these fractures is
surgical. In B1 fractures, open reduction and
internal fixation (ORIF) with one or more plates
is generally advocated.2 Locking plates (LPs) are
widely used for the treatment of periprosthetic
femoral fractures (PPFFs). They provide theoret-
ical advantages, by ensuring angular stability of
the screw to the plate, using monocortical screws
for fixation about the femoral component, and
minimizing soft-tissue trauma.3 However, previ-
ous reports have shown differing results regarding
rates of fracture union,4-6 when LPs are compared
with conventional plates (CPs). While femoral
component revision may be advocated in spe-
cific cases of B1 fractures,7 it is regarded as the
treatment of choice for fractures around a loose
femoral component.8 To our knowledge, no pre-
vious study has compared the outcomes following
the use of cemented, cementless monoblock, or
cementless modular revision components in the
treatment of PPFF associated with a loose femo-
ral component. Interprosthetic femoral fractures
(IPFFs), fractures between ipsilateral total knee
and hip arthroplasties, have been reported to have
similar outcomes to periprosthetic fractures in
1447
1448 G. Chatziagorou, H. Lindahl, J. Kärrholm 

Fig. 1

Flowchart showing the exclusion criteria of the study. PPFF, periprosthetic femoral fracture;
THA, total hip arthroplasty.

patients without a total knee arthroplasty (TKA) distal to the
fracture site. However, these studies also included results from
patients with hemiarthroplasties, primary total hip arthroplast-
ies (THAs), and revision THAs.6,9
The aim of this registry study was to: 1) describe the surgi-
cal treatment and outcomes of Vancouver type B fractures in
Sweden; 2) investigate if IPFFs had poorer outcomes than non-­
IPFFs; 3) compare outcomes in Vancouver type B1 fractures
treated by either CPs or LPs; and 4) study outcomes, where
revision of the femoral component was used for the treatment of
type B2 and B3 fractures. Any reoperation following the treat-
ment of PPFF was the primary outcome measure.
Patients and Methods
The study was approved by the Central Ethical Review Board
in Gothenburg (entry number: 198-12; date: 5 April 2012).
Data collection. The study included all primary THAs operated
in Sweden since 1979, and subsequently underwent any further
surgery due to PPFF between 2001 and 2011. This cohort was
followed up until 31 December 2013 or until any further reop-
eration, death, or emigration, giving a minimum of two years’
follow-up (maximum follow-up 13 years). The study group
was identified primarily from data in the Swedish Hip Arthro-
plasty Register (SHAR),10 where information regarding patient
demographics, component characteristics, and procedure
details are registered prospectively. In order to detect PPFFs
surgically treated between 2001 and 2011, and not registered
in the SHAR, data-linking with the National Patient Register
(NPR) was performed.11 A second linkage between SHAR and
NPR was performed in order to detect any type of reoperation
following a PPFF, performed between 2001 and 2013, and reg-
istered only in the NPR. The linkage was based on KVÅ treat-
ment codes (Klassifikation av vårdåtgärder, Classification of
Health Care Procedures),12 corresponding to the OPCS (Classi-
fication of Intervention and Procedures) of the National Health
Service (NHS) of the United Kingdom.13 Any kind of surgical
procedure in the hip/femur was derived from the NPR database.
All medical records of cases involving a hip prosthesis and a
fracture reported only to the NPR were collected and analyzed.
Medical records of patients undergoing any reoperation and
registered to SHAR, were routinely reviewed by the personnel
at SHAR when their data were entered. Information regarding

Follow us @BoneJointJ THE BONE & JOINT JOURNAL

 Surgical treatment of Vancouver type B periprosthetic femoral fractures 1449
the classification of a PPFF, the surgical procedure, the mecha-
nism of injury, discharge destination, and prescribed postoper-
ative weightbearing status was recorded. Details regarding the
transfusion requirements, fracture classification, and its valida-
tion, have been described previously.11 Data regarding the reop-
erations were obtained from SHAR or from the medical records
of cases identified after the second cross referencing with NPR.
Interprosthetic fractures were identified either by information
from the medical records or from data-linking with the Swedish
Knee Arthroplasty Register (SKAR).14
Exclusions and definitions. Hip resurfacing arthroplasties and
intraoperative fractures were excluded from the study. Femoral
shaft perforations, cases with ongoing infection or with diag-
nosis of malignancy, fractures during knee surgery, and further
exclusions are listed in Figure 1 and Table I. Reoperation was
defined as any further surgical intervention related to the index
hip arthroplasty, irrespective of whether the prosthesis or parts
of it were exchanged, extracted, or left untouched. Reopera-
tions for the treatment of PPFF were divided into three groups:
the ORIF group (with plate, nail, screw, cerclage, or Kirschner
wire); the femoral component revision group; and the ORIF plus
femoral component revision group. In the ORIF group, no revi-
sion surgery of the femoral component was performed. Femoral
component revision was defined as any kind of revision of the
component and or any of its parts, but without the use of ORIF
(except from cerclage wiring). Cases treated by cement-in-­
cement revision were regarded as a femoral component revision,
even when the original component was reinserted. The use of
bone graft was recorded and divided into use of strut allograft
or any other type of bone grafting. Any reoperation due to non­
union, refracture, fixation failure, hip dislocation, femoral com-
ponent loosening, infection, pain, and other technical reasons
was regarded as a reoperation related to the surgical treatment
of PPFF. For the purposes of data analysis, nonunion, refracture,
and fixation failure were merged into one group and referred to
as ‘nonunion’. Amputation caused by arterial insufficiency and
reoperations due to acetabular component loosening were not
considered to be related to the previously treated PPFF. Patients
prescribed partial weightbearing and full weightbearing were
combined into one group (‘allowed weightbearing’), as partial
weightbearing could not accurately be accomplished.15
Subgroups of the study material. The statistical analysis was
performed in two steps. First, we analyzed all first-time reoper-
ations including all subcategories (B1, B2, B3). Thereafter, we
separately studied the surgical treatment of fractures around a
well-fixed femoral component, and those fractures close to a
loose component. In type B1 fractures not treated with femo-
ral revision, we compared CPs with LPs. This was done after
further excluding of IPFFs, double plating, unknown plates,
and additional procedures (Fig. 2). In the combined group
of Vancouver type B2 and B3 fractures treated with femoral
component revision (with or without ORIF), we evaluated the
influence of femoral component design according to the three
principal groups (cementless modular, cementless monoblock,
and cemented). Interprosthetic fractures were also excluded in
this subanalysis (Fig. 2). In order to have a more comparable
and homogeneous material within the revision categories, we
included only femoral components longer than 150 mm (with
VOL. 101-B, No. 10, OCTOBER 2019
Table I. Other exclusion criteria (n = 93)
Exclusion criteria Cases, n
Intraoperative fracture 57
Active deep infection at the time of PPFF 20
Perforation only 10
Iatrogenic fracture during TKA surgery 5
PPFF due to sawing (noniatrogenic) 1
PPFF, periprosthetic femoral fracture; TKA, total knee arthroplasty.
known length), and prostheses used in more than ten cases dur-
ing the whole study period (Table II). We excluded four MP
modular cementless femoral components because cemented
fixation was used for their distal part. For both subgroups (B1
and B2/B3), interprosthetic fractures were not included. Further
exclusion criteria are shown in Figure 2 and Table II.
Statistical analysis. Statistical calculations were done using
IBM SPSS statistics version 25 (IBM Corp., Armonk, New
York). All reoperation rates referred in this study were calcu-
lated based on PPFF-related reoperations. Cox regression anal-
ysis for the estimation of relative risk had as primary outcome
any PPFF-related reoperation. The following were censored:
cases with reoperation due to acetabular component loosening
or arterial insufficiency, patients who died without any reopera-
tion, or those who had not undergone further reoperation before
the end of 2013. Reoperation rates and proportional compari-
son between groups was performed with chi-squared test and
Fisher’s exact test. Comparison of means was done using Stu-
dent’s t-test and, when applicable, one-way analysis of vari-
ance (ANOVA) and Tukey test. p-values were two-sided with
a significance level < 0.05, and 95% confidence intervals (CIs)
were calculated. The STROBE (Strengthening the Reporting of
Observational Studies in Epidemiology) checklist was used.
Results
Study material: demographics. Between 2001 and 2011, 3220
reoperations due to PPFF (2940 patients) were registered. After
the exclusions listed in Figure 1 and Table I, 1381 Vancouver
type B fractures were left for analysis (1373 patients). At the
time of primary THA, patients with a fracture close to a well-
fixed femoral component were older than those with a fracture
around a loose component (B2: p = 0.004; B3: p = 0.002; Table
III). The most common diagnosis at the time of primary THA
was primary osteoarthritis (OA) and the most common cause of
fracture was fall from standing height (Table III). Type B frac-
tures occurred earlier following primary THA if the patient´s
primary diagnosis was avascular necrosis of the femoral head
(mean 7.7 years, 95% CI 5.9 to 9.4) or hip fracture (mean 6.2
years, 95% CI 5.3 to 7.2) compared with those operated due to
primary OA (mean 10.4 years, 95% CI 10.0 to 10.8) or inflam-
matory arthritis (mean 12.1 years, 95% CI 10.6 to 13.7). The
mean time between primary THA and PPFF (Table III) was
greatest for B3 fractures (p < 0.001 vs B1; p = 0.001 vs B2) and
shortest for type B1 fractures (p = 0.001 vs B2). The distribu-
tion of fracture types differed between the sexes, with relatively
more type B2 fractures among male patients and more type B1
and B3 among female patients (Table III). Significant difference
was noticed only between B1 and B2 fractures (p = 0.012). At
1450 G. Chatziagorou, H. Lindahl, J. Kärrholm 

Fig. 2

Flowchart showing cases excluded from the subanalysis of B1 and B2/B3 fractures, respectively. ORIF, open reduction and internal fixation; IPFF,
interprosthetic femoral fractures.

Table II. Excluded revision femoral components from the subanalysis of B2 and B3 fractures

Exclusion criteria Manufacturer Cases, n (n= 57)

Short* or unknown length 30
Exeter Polished Stryker, Mahwah, New Jersey 15
CPT Zimmer Biomet, Warsaw, Indiana 7
Lubinus SP II Link, Hamburg, Germany 6
Spectron EF Primary Smith & Nephew, Memphis, Tennessee 2
Components used in < 10 procedures during 2001 to 2011 23
METS Stanmore, Elstree, United Kingdom 4
Restoration Straight Stryker, Mahwah, New Jersey 4
ZMR Zimmer Biomet, Warsaw, Indiana 3
BiMetric Biomet, Bridgend, United Kingdom 3
Corail DePuy Synthes, Warsaw, Indiana 2
Hyperion Biomet, Berlin, Germany 2
MS 30 Zimmer, Winterthur, Switzerland 1
ScanHip Revision MitAB, Sjöbo, Sweden 1
Meridian Stryker/Howmedica, Mahwah, New 1
Jersey
CLS Zimmer, Germany 1
ABG Stryker, Montreux, Switzerland 1
Cementless components with distal cemented fixation 4
MP revision Link, Hamburg, Germany 4
*Defined as 150 mm or shorter

the time of PPFF, the mean age of the male population with type population with B1 (80.0 years, 95% CI 78.4 to 81.7; p = 0.008)
B1 (76.6 years, 95% CI 74.5 to 78.6) or type B2 fracture (76.9 or B2 fracture (79.4 years, 95% CI 78.4 to 80.5; p < 0.001),
years, 95% CI 76.0 to 77.8) was lower than that of the female respectively. Vancouver type B3 fractures occurred at similar
Follow us @BoneJointJ THE BONE & JOINT JOURNAL
 Surgical treatment of Vancouver type B periprosthetic femoral fractures 1451

Table III. Demographics, treatment, and outcome of periprosthetic femoral fractures (PPFFs) in relation to fracture type

Characteristic Type of PPFF (Vancouver classification)

B1 B2 B3 Total
All fractures, n 317 989 75 1381
Interprosthetic fractures, n (%) 18 (5.7) 74 (7.5) 5 (6.7) 97 (7.0)
Male sex,* n (%) 146 (46.1) 536 (54.2) 35 (46.7) 717 (51.9)
Cemented primary component, n (%) 288 (90.9) 859 (86.7) 75 (100) 1222 (88.5)
Diagnosis at primary THA, n (%)
Primary osteoarthritis 212 (66.9) 725 (73.3) 54 (72.0) 991 (71.8)
Hip fracture 55 (17.3) 130 (13.2) 7 (9.3) 192 (13.9)
Inflammatory arthritis 23 (7.3) 60 (6.1) 8 (10.7) 91 (6.6)
Idiopathic avascular necrosis 16 (5.0) 40 (4.0) 2 (2.7) 58 (4.2)
Other 11 (3.5) 34 (3.4) 4 (5.3) 49 (3.5)
Mechanism of injury, n (%)
Fall from standing height 271 (85.5) 830 (83.9) 61 (81.4) 1162 (84.1)
No known trauma 29 (9.1) 115 (11.6) 12 (16.0) 156 (11.3)
Higher energy 14 (4.4) 39 (4.0) 1 (1.3) 54 (3.9)
Unknown reason 3 (0.9) 5 (0.5) 1 (1.3) 9 (0.7)
Mean age, yrs (95% CI)
At primary THA* 70.0 (68.7 to 71.3) 68.0 (67.3 to 68.6) 65.5 (63.3 to 67.8) 68.3 (67.7 to 68.9)
At PPFF 78.4 (77.1 to 79.7) 78.1 (77.4 to 78.7) 78.8 (76.6 to 81.0) 78.2 (77.6 to 78.8)
Mean time to PPFF,* yrs (95% CI) 8.4 (7.7 to 9.1) 10.1 (9.6 to 10.6) 13.3 (11.9 to 14.7) 9.9 (9.5 to 10.3)
Treatment, n (%)
ORIF 287 (90.5) 135 (13.6) 1 (1.3) 423 (30.6)
ORIF + femoral component revision 6 (1.9) 144 (14.6) 24 (32.0) 174 (12.6)
Femoral component revision only 24 (7.6) 710 (71.8) 50 (66.7) 784 (56.8)
Discharged to, n (%)
Other unit 208 (65.6) 644 (65.1) 48 (64.0) 900 (65.2)
Home 71 (22.4) 255 (25.8) 20 (26.7) 346 (25.0)
Died 8 (2.5) 23 (2.3) 2 (2.6) 33 (2.4)
Unknown 30 (9.5) 67 (6.8) 5 (6.7) 102 (7.4)
Reoperation until 2013, n (%)
Related to PPFF* 70 (22.1) 154 (15.6) 15 (20.0) 239 (17.3)
All causes 74 (23.3) 167 (16.9) 16 (21.3) 257 (18.6)
Reoperation registered in SHAR,† n (%) 61 (82.4) 136 (81.4) 14 (87.5) 211 (82.1)
Deceased by 31 December 2013, n (%) 158 (49.8) 507 (51.2) 49 (65.3) 714 (51.7)
Mean time to any reoperation, yrs (95% CI) 1.2 (0.8 to 1.5) 1.4 (1.1 to 1.7) 1.6 (0.5 to 2.8) 1.3 (1.1 to 1.5)
Mean time of follow-up, yrs (95% CI) 3.4 (3.1 to 3.8) 3.8 (3.6 to 3.9) 4.1 (3.3 to 4.9) 3.7 (3.5 to 3.9)
*There was a statistically significant difference between the groups studied
†The percentage is calculated from all reoperations (all causes) within each Vancouver category
THA, total hip arthroplasty; CI, confidence interval; ORIF, open reduction and internal fixation; SHAR, Swedish Hip Arthroplasty Register

age in male and female patients. Only 25% of all patients were
discharged directly to home following surgical treatment of
PPFF. Approximately half of the study group had died at the
end of the study period in 2013.
Validation of registration at the SHAR and reoperation
rate. At the endpoint of the study in 2013, 257 patients (257
THAs) had undergone a further reoperation, with 211 of them
(82.1%) being reported to the SHAR. The registration rate was
higher (103/108, 95.4%) for reoperations, which include revi-
sion of the femoral component, and lower (108/149, 72.5%)
for reoperations other than femoral component revision. The
reoperation rate for patients with IPFF was higher (29/97, 30%)
than for those without an ipsilateral knee prosthesis (210/1284,
16%; p = 0.001). Fractures around a well-fixed femoral compo-
nent had significantly higher reoperation rate than B2 fractures
(p = 0.010; Table III). This difference was still significant after
excluding the interprosthetic fractures (p = 0.002). The reasons
VOL. 101-B, No. 10, OCTOBER 2019
for reoperation in relation to Vancouver categories are shown
in Figure 3. To evaluate the influence of a TKA distal to the
fracture, we used a Cox regression model with adjustment for:
Vancouver category (B1 or B2/B3), sex, age at PPFF, and year
of operation due to PPFF. The population studied was only
patients with primary OA and cemented primary femoral com-
ponent (880 PPFFs with 139 reoperations). Overall, the risk for
reoperation was two times higher for Type B PPFFs in the pres-
ence of an ipsilateral TKA (Table IV).
Vancouver type B1 fractures: conventional plate versus
locking plate. A total of 30 type B1 fractures were treated by
femoral component revision (with or without ORIF) and with-
out a significantly different reoperation rate (five cases, 17%),
compared with those treated by ORIF only (65 cases, 23%);
even after excluding the IPFFs (19% and 22%, respectively).
Within the ORIF group, one B1 fracture was treated with
intramedullary nail fixation (‘docking nail method’) and seven
1452 G. Chatziagorou, H. Lindahl, J. Kärrholm 

Fig. 3

Chart showing reoperation rates (%) in relation to Vancouver category and reason for reoperation. All peri-
prosthetic femoral fractures (1381 Vancouver type B fractures) and all reoperations (257 cases) are included.

Table IV. Cox-regression analysis with risk factors, hazard ratios (HRs), and 95% confidence intervals (CIs) for re-
operation after surgical treatment of Vancouver type B periprosthetic femoral fracture (PPFF). Only cases with the
diagnosis of primary osteoarthritis and cemented femoral fixation at primary total hip arthroplasty are included

Characteristic Unadjusted Adjusted

HR (95% CI) p-value HR (95% CI) p-value
TKA ipsilaterally 0.019 0.009
No (reference) 1 1
Yes (IPFF) 2.0 (1.1 to 3.5) 2.2 (1.2 to 3.8)
Type of PPFF 0.026 0.025
B2 or B3 (reference) 1 1
B1 1.5 (1.1 to 2.2) 1.5 (1.1 to 2.2)
Sex 0.388 0.317
Female (reference) 1 1
Male 0.9 (0.6 to 1.2) 0.8 (0.6 to 1.2)
Mean age at PPFF 0.977 (0.959 to 0.996) 0.016 0.973 (0.955 to 0.992) 0.005
Year of operation for PPFF 0.99 (0.94 to 1.05) 0.737 0.99 (0.94 to 1.05) 0.699
IPFF, interprosthetic femoral fracture

fractures were fixed with cerclage only. The rest of the 279 B1
fractures were treated with either a single plate (244 cases) or
with two plates (35 cases). In 14 cases with single plate fix-
ation, additional procedures were also required (one Kirschner
wire, two bone cement augmentation, seven bone graft, four
revisions of the modular femoral head or the acetabular com-
ponent). Finally, following exclusions, 212 B1 fractures
treated with either a CP (124) or a LP (88) were left for anal-
ysis (Fig. 2). The first LP in this treatment group analysis was
utilized in 2005 (Fig. 4). More female patients and a shorter
follow-up were recorded in the LP group (Table V). Conven-
tional plating had a higher reoperation rate (25.8%) compared
with the locking plate group (19.3%), but this was not statisti-
cally significant (p = 0.322). A greater proportion of patients
needed femoral component revision at the first reoperation fol-
lowing the PPFF in the CP group (54.5%), compared with the
LP group (16.7%, p = 0.016). The mean time to reoperation after
PPFF was 1.1 years (95% CI 0.7 to 1.4), without significant
difference between CP and LP. Reosteosynthesis was applied
for the majority of the complications, not treated with femoral
component revision (CP: eight cases; LP: 13 cases). The most
common reason for reosteosynthesis was nonunion. The indica-
tions for the first reoperation are illustrated in Table VI.
Vancouver type B2-B3 fractures and revision femoral com-
ponent categories. A significantly greater reoperation rate
(p = 0.015) was noticed for fractures treated only with ORIF
(28/127 cases, 22%), compared with those treated with femoral
component revision in isolation or in combination with other

Follow us @BoneJointJ THE BONE & JOINT JOURNAL

 Surgical treatment of Vancouver type B periprosthetic femoral fractures 1453

Fig. 4

Chart showing the number of Vancouver type B1 fractures treated with the use of either a conventional or a
locking plate. Excluded cases are illustrated in Figure 2.

Table V. Conventional versus locking plate fixation in Vancouver type B1 fractures. Excluded cases are illustrated
in Figure 2

Characteristic Type of plate fixation p-value

Conventional Locking Total
All fractures, n 124 88 212
Male sex, n (%) 69 (55.6) 35 (39.8) 104 (49.1) 0.026*
Cemented primary component, n (%) 113 (91.1) 82 (93.2) 195 (92.0) 0.798*
Diagnosis of primary OA, n (%) 91 (73.4) 59 (67.0) 150 (70.7) 0.359*
Mean age, yrs (95% CI)
At primary THA 69.9 (68.0 to 71.9) 73.1 (70.7 to 75.5) 71.2 (69.7 to 72.7) 0.043†
At PPFF 77.9 (75.9 to 79.9) 80.9 (78.6 to 83.2) 79.2 (77.6 to 80.7) 0.054†
Weightbearing,‡ n (%)
Allowed 73 (59.3) 62 (70.5) 135 (64.0) 0.111*
Not allowed 50 (40.7) 26 (29.5) 76 (36.0) 0.111*
Reoperation related to PPFF, n (%) 32 (25.8) 17 (19.3) 49 (23.1) 0.322*
Deceased by 31 December 2013, n (%) 71 (57.3) 36 (40.9) 107 (50.5) 0.026*
Mean time of follow-up, yrs (95% CI) 3.8 (3.2 to 4.4) 2.6 (2.2 to 3.0) 3.3 (2.9 to 3.7) 0.001†
*Fisher’s exact test
†Student’s t-test
‡There was one case with lack of information regarding postoperative weightbearing in the conventional plating
group
OA, osteoarthritis; CI, confidence interval; THA, total hip arthroplasty; PPFF, periprosthetic femoral fracture

procedures (116/858, 13.5%). Strut allograft was used in only
one B2 fracture, which was later reoperated due to femoral com-
ponent loosening. Three out of seven B3 fractures treated with
strut allograft underwent a further surgery due to dislocation,
refracture, or heterotopic ossification, respectively. Other types
of bone graft were used in 55 cases of 914 type B2 fractures,
and in 21 cases of 63 type B3. The reoperation rate (18.4%) was
not statistically different (p = 0.305) in these cases compared
with those treated without bone grafting (14%).
After the exclusion criteria described were applied (Fig. 2,
Table II), 801 type B2 and B3 fractures, treated with femoral
component revision (with or without ORIF), were available
for analysis (Table VII). There were demographic differences
between the three femoral component categories (cementless
modular, cementless monoblock, and cemented). The cemented
group were older patients with a higher proportion of females
and with more B3 fractures. Fewer of the patients in the cement-
less modular group had died by the end of the study period,
which was probably explained by a shorter follow-up period.
Almost half of the modular cementless femoral revision com-
ponents (48%) were inserted between 2009 and 2011 (Fig. 5).
The mean time to reoperation was 1.4 years (95% CI 1 to 1.7)
for all 108 PPFF-related reoperations and did not differ sta-
tistically between the femoral component revision categories.

VOL. 101-B, No. 10, OCTOBER 2019

1454 G. Chatziagorou, H. Lindahl, J. Kärrholm 

Table VI. Reasons for reoperation following surgical treatment of Vancouver type B1 fractures with either
a conventional (n = 124) or a locking plate (n = 88)

Conventional plate: all Locking plate: all

reoperations (revisions) reoperations (revisions)
Nonunion 17 (7) 15 (2)
Loosening of femoral component/total 8 (8) 1 (1)
Infection 5 (2) 1 (0)
Hip dislocation 1 (1) 0 (0)
Technical reasons 1 (0) 0 (0)
Loosening acetabular component 1 (0) 0 (0)
Arterial insufficiency 0 (0) 1 (0)
Total 33 (18) 18 (3)

Table VII. Demographics and outcome after treatment of B2 and B3 fractures with femoral component revision, alone or with open reduction and
internal fixation. Excluded cases are illustrated in Figure 2 and Table II

Characteristic Cementless Cementless Cemented All femoral p-value

modular monoblock components
Fracture type, B2, n (%) 401 (94.4) 97 (94.2) 239 (87.5) 737 (92.0) Modular vs cemented:
0.002
Male sex, n (%) 255 (60.0) 59 (57.3) 138 (50.5) 452 (56.4) Modular vs cemented:
0.015
Cemented primary femoral 367 (86.4) 92 (89.3) 242 (88.6) 701 (87.5)
component n (%)
Diagnosis of primary OA, n (%) 329 (77.4) 75 (72.8) 197 (72.2) 601 (75.0)
Mean age in years (95% CI)
At primary THA 66.7 (65.8 to 67.7) 66.5 (64.3 to 68.7) 69.0 (67.9 to 70.2) 67.5 (66.8 to 68.2) Modular vs cemented:
0.009; monoblock vs
cemented: 0.070
At PPFF 77.2 (76.2 to 78.2) 77.2 (75.1 to 79.3) 80.0 (78.8 to 81.3) 78.1 (77.4 to 78.9) Modular vs cemented:
0.001; monoblock vs
cemented: 0.046
Reason for reoperation, n (%)
All reasons* 69 (16.2) 14 (13.6) 36 (13.2) 119 (14.9)
Nonunion 9 (2.1) 3 (2.9) 16 (5.9) 28 (3.5)
Hip dislocation 23 (5.4) 7 (6.8) 10 (3.7) 40 (5.0)
Infection 16 (3.8) 1 (1.0) 5 (1.8) 22 (2.7)
Loosening of femoral 10 (2.4) 1 (1.0) 1 (0.4) 12 (1.5)
component/total
Technical reasons/pain 3 (0.7) 1 (1.0) 2 (0.7) 6 (0.7)
Weightbearing, n (%) Modular vs cemented:
0.002; monoblock vs
cemented: 0.021†
Not allowed 51 (12.0) 13 (12.6) 13 (4.8) 77 (9.6)
Allowed 166 (86.1) 89 (86.4) 246 (90.1) 701 (87.5)
Unknown 8 (1.9) 1 (1.0) 14 (5.1) 23 (2.9)
Mean time of follow-up, yrs (95% CI) 3.6 (3.3 to 3.8) 4.9 (4.3 to 5.6) 4.1 (3.7 to 4.5) 3.9 (3.7 to 4.2) Modular vs monoblock:
< 0.001; monoblock vs
cemented: 0.049
Deceased (31 December 2013), n (%) 169 (39.8) 58 (56.3) 176 (64.5) 403 (50.3) Modular vs monoblock:
0.003; modular vs
cemented: < 0.001
Total 425 103 273 801
*Reoperation due to acetabular component loosening and arterial insufficiency are referred in the text
†After exclusion of missing cases
OA, osteoarthritis; CI, confidence interval

Reoperation due to acetabular component loosening was noted
only in cementless modular components (three cases with MP
femoral prosthesis and five cases with Revitan prosthesis). One
patient treated with a Wagner prosthesis and two PPFFs treated
with cemented femoral components (Spectron and CPT),
underwent amputation above the knee due to arterial insuffi-
ciency. All other reasons for reoperation are illustrated in Table
VII. Femoral components used in each category are described
in Table VIII. The reoperation rate did not statistically differ
between the three femoral component categories.

Follow us @BoneJointJ THE BONE & JOINT JOURNAL

 Surgical treatment of Vancouver type B periprosthetic femoral fractures 1455

Fig. 5

Number of B2 and B3 periprosthetic fractures treated with femoral component revision, alone or in combi-
nation with open reduction and internal fixation (801 cases). Excluded cases are illustrated in Figure 2 and
Table II.

Table VIII. Revision femoral components used for the treatment of B2 and B3 fractures, and included in the subanalysis

Type of revision femoral component Manufacturer PPFFs (reoperations*), n Reoperation rate, %

Cementless modular 425 (61) 14.4
MP revision Link, Hamburg, Germany 215 (43) 20.0
Revitan Zimmer, Warsaw, Indiana 114 (13) 11.4
Restoration Stryker, Mahwah, New Jersey 96 (5) 5.2
Cementless monoblock 103 (13) 12.6
Wagner Smith & Nephew, Rotkreuz, Switzerland 103 (13) 12.6
Cemented 273 (34) 12.5
Lubinus SP II Link, Hamburg, Germany 103 (19) 18.4
Exeter Polished Stryker, Mahwah, New Jersey 79 (7) 8.9
Spectron Smith & Nephew, Memphis, Tennessee 44 (3) 6.8
CPT Zimmer Biomet, Warsaw, Indiana 19 (1) 5.2
Charnley DePuy, Leeds, United Kingdom 16 (3) 18.8
RX 90-S Biomet, Warsaw, Indiana 12 (1) 8.3
Total 801 (108) 13.5
*Reoperations related to periprosthetic femoral fractures (PPFFs) as associated with femoral prosthesis

Discussion manner. We observed that the LP group had a higher percent-

Postoperative periprosthetic fracture around or close to a pri-
mary THA is a relatively rare complication11,16,17 and its treat-
ment is often individualized. Large cohorts of patients are
needed in order to draw more robust conclusions. National
registries have considerable variability on types and methods
of data collection, which implies a potential risk of bias when
interpreting the results. The SHAR is the only arthroplasty reg-
ister that has recorded all types of reoperations since 1979. The
ability to link data with the NPR, based on the personal iden-
tity number, allows for very high accuracy in data collection
and gives the opportunity for continuous data validation and
improvement to SHAR’s database. The comparison between
patients treated with CPs and LPs for the treatment of type
B1 fractures was performed in a meticulously and systematic
age of women and older patients compared with the CP group
(Table V). This may represent a selection bias, where surgeons
preferred fixation with LP for fractures in osteoporotic bones.
Furthermore, LPs were not used for the treatment of B1 frac-
tures before 2005 (Fig. 4), resulting in a shorter follow-up
(Table V). Previous studies have reported increased daily activ-
ity levels18 and higher risk for aseptic loosening19 with type B
fractures in men with THA.20
A previous biomechanical study of plates and screws on a
cemented THA model21 showed that crack formation within
the cement mantle was associated with the position of the
screw used, rather than the design of the plate. Our finding of
a higher rate of femoral revisions following fixation with a CP
(55%) should be interpreted with caution. We did not observe

VOL. 101-B, No. 10, OCTOBER 2019

1456 G. Chatziagorou, H. Lindahl, J. Kärrholm 
significantly higher reoperation rates of fractures treated with
CP (26%) compared with those fixed with a LP (19%). Pre-
viously, authors have tried to answer this question, using the
rate of nonunion as the primary outcome measure. Moore et al5
showed no significant difference between the groups, Dehghan
et al4 reported a higher nonunion rate for the LPs, and ­Stoffel
et al6 noticed that LPs had 12 times lower relative risk for non-
union. However, the reoperation rate was not significantly dif-
ferent in the two latter studies. Our results, therefore, give no
firm conclusion as to the optimum type plate that should be
used for ORIF of a Vancouver type B1 fracture.
Fractures close to a loose femoral prosthesis were analyzed
as one category. The distinction between a B2 and a B3 fracture
is not always simple, as the assessment of bone stock may dif-
fer between surgeons but both fracture types demand revision
of the component. We identified only eight cases treated with
supplementary strut allografts among 915 type B2 and 70 type
B3 fractures. This is very low compared with previous reports,
where strut allografts were used even for the treatment of type
B1 fractures.22,23 The comparison of modular or mono­block
cementless femoral components, and cemented components did
not result in significantly different reoperation rates. However,
cemented prostheses had a tendency to be chosen in elderly,
female patients with type B3 fractures (Table VII), as often
reported in the literature.22 Throughout the study period, we
note a steady increase in the use of cementless modular femoral
components (Fig. 5). These femoral revision components had the
shortest follow-up compared with cementless monoblock and
cemented femoral prostheses (Table VII). Moreta et al24 com-
pared 19 Wagner femoral components with 24 Modular-Plus
cementless modular components and found significantly more
infections when the monoblock revision prostheses was used.
Previous studies report no difference in the reoperation rate
between cemented and cementless fixation of the revision com-
ponent.25,26 Other authors have reported that cemented revision
components had the highest re-revision rate (19%), compared
with fully (13%) and proximally coated (11%) revision cement-
less prostheses, in an evaluation of 118 Vancouver type B frac-
tures.22 However, proximally coated femoral prostheses had
poorer outcomes due to a much higher rate of complications
(64%). We noticed a higher reoperation rate of MP and Lubinus
SPII prostheses, in comparison with the Restoration femoral
component (Table VIII), which needs further investigation.
Our study supports the accepted axiom that a fracture around
a loose femoral component should be treated by revision. Ιn
B1 fractures, the choice between femoral revision and ORIF
did not influence the outcome, as reported in a previous study.7
The relatively small number of primary cementless stems and
diagnoses other than primary OA made us restrict the study of
B fractures to those occurring close to a cemented component
and only in patients with primary OA (Table IV). To our knowl-
edge, this is the first clinical study to estimate the hazard ratio
regarding the influence of an ipsilateral knee arthroplasty to the
outcome of Vancouver B fractures.
As with previous reports,17,26 B1 fractures had a poorer out-
come when compared with B2 and B3 fractures. Several hypoth-
eses may be considered to explain this finding. Some B1 fractures
were probably in reality B2 fractures and therefore erroneously
Follow us @BoneJointJ
assessed to have a well-fixed femoral component. B2 fractures
had poorer outcomes in our data when treated with ORIF only.
This ‘misclassification’ of B1 fractures could contribute to a
poorer outcome in this group. Another possible explanation is
that the bone-cement interface or the bone-femoral component
interface in cementless prostheses may be partially damaged or
interrupted in the majority of B1 fractures. Although the sur-
geon, by visual inspection and manual testing, may perceive
the component to be securely fixed, the quality of the fixation
might be impaired, making the component susceptible to future
loosening. This explanation could be supported by the finding
that the second most common reason for reoperation after treat-
ment of a B1 fracture was aseptic femoral loosening (Fig. 3),
and femoral component revision was performed in many cases
with nonunion (Table VI). Moreover, even if the fracture did
not affect the stability of the femoral component, there may
be increased risk for the cement-component interface being
adversely affected by the screws used to fix the plate. Insertion
of a longer femoral component bypassing the fracture might
function as ‘intramedullary fixation’, which might be superior
to extramedullary fixation with ORIF by a plate. Pavlou et al7
studied this hypothesis and they found a shorter time to union,
but no difference in union rate, when comparing femoral compo-
nent revision with plate fixation associated with transverse frac-
tures close to the tip of the primary femoral prosthesis. In our
results, the treatment of a B fracture with intramedullary fixation
(longer femoral component revision) resulted in a higher rate
of immediate postoperative weightbearing (87.5%, Table VII),
compared with those patients treated with plate fixation alone
(64%, Table V). This difference may reflect higher stability of
fracture fixation with femoral revision, and consequently better
conditions for fracture healing. Finally, the location and the type
of the fracture may also play an important role on the outcome
of the treatment with plate fixation. A transverse or short oblique
fracture at the tip of the component is more difficult to stabilize
with an extramedullary fixation than with an intramedullary fix-
ation (femoral component revision). We did not analyze radio-
graphs in this study, and we considered the information from the
medical charts, regarding radiological outcomes, as unreliable.
Possibly, there might have been a higher proportion of fractures
close to the tip of the component in the B1 group that required a
further reoperation, but this requires further investigation.
Our study has several limitations. Reoperation rate is a crude
measure of the outcome of fracture treatment. Previous stud-
ies have shown that patients with radiological femoral loos-
ening or nonunion following PPFF treatment are not always
re-revised.17,22,26 This could be explained by lack of symptoms,
patient comorbidities, and the risk for further complications
with further intervention. The recording of patient related out-
come measurements (PROMs) for reoperations was initiated in
SHAR in 2017, and could therefore not be used in this eval-
uation. Our study also lacks a radiological evaluation. Infor-
mation regarding the type of fracture was extracted from the
case records only. The optimum would be to combine infor-
mation from both the radiographs and the medical records, but
this was not regarded as feasible in such a large study. Further,
the classification of fractures based on medical records is not
necessarily inferior to a radiological assessment of the fracture
THE BONE & JOINT JOURNAL
 Surgical treatment of Vancouver type B periprosthetic femoral fractures 1457
types. Several studies emphasized the discrepancy in evaluation
of type B fractures based on preoperative radiographs compared
to intraoperative findings.26,27 In the current study, classification
was chiefly based on the operative findings, where information
about component loosening was derived. The classification
process was previously validated11 and good agreement was
observed corresponding to previous reports of the Vancouver
classification.28 The role of locking attachment plates (LAP)
compared with the use of cerclage wiring in combination with
plate fixation was not investigated. Biomechanical studies have
shown better results with LAPs.29,30 Such an analysis in the
ORIF group could not be performed in our study due to small
numbers in some of the treatment groups.
The data-linking between the SHAR and the NPR, for both
PPFFs and reoperations, is one of the methodological strengths
of this research. In the first cross-referencing, we looked for
first-time reoperations due to PPFF missing from the SHAR. In
the second cross-referencing, we traced any unrecorded reop-
erations following a PPFF, including those reoperations whose
index PPFF had not been reported to the SHAR. Furthermore,
we were able to record the presence of a TKA, ipsilaterally to
the hip prosthesis, based on data from both the medical records
and the SKAR. This cross-referencing between multiple data
sources ensured a good data quality and a high external valid-
ity. The follow-up in the present study was long enough (mean
3.7 years, Table III) to capture the great majority (if not all) of
the failures after PPFF, since the mean time to reoperation was
1.3 years. We feel that the possibility to exclude many factors
that may had influenced outcomes is a methodological strength.
Hemiarthroplasties were not included, and none of the prosthe-
ses studied had been revised or undergone any other reoperation
prior to the PPFF. This highlights the unique strengths of the
SHAR and data collection in Sweden. The analysis of plate fix-
ation in B1 fractures, was performed in a group where only one
plate was used and no other additional surgery was performed.
Interprosthetic fractures, according to our findings, had a neg-
ative influence on the outcome of Vancouver type B fractures,
and were excluded from both subanalyses (B1 and B2/B3),
which gives greater clarity to outcomes.
In conclusion, a poorer outcome was observed following
surgical treatment of Vancouver type B1 fractures. Fractures
close to a loose prosthesis should be treated with femoral com-
ponent revision. The type of revision component utilized did
not influence the outcome. CPs and LPs had similar reoperation
rates after treatment of B1 fractures. However, a higher rate of
reoperations requiring femoral component revision was noticed
in B1 fractures treated with CPs. Primary OA patients with a
cemented primary femoral component who were treated surgi-
cally due to Vancouver type B fracture had a two times greater
risk for subsequent reoperation if they had an ipsilateral TKA.
Take home message
- Vancouver type B fractures had higher risk for reoperation
when a total knee arthroplasty was present ipsilaterally.
- No difference in the reoperation rate between locked and conventional
plating for the treatment of Vancouver type B1 fractures was observed.
- Cemented, cementless monoblock, and cementless modular femoral
components for the treatment of Vancouver type B2 and B3 fractures had
similar outcomes.
VOL. 101-B, No. 10, OCTOBER 2019
References
1. Duncan CP, Masri BA. Fractures of the femur after hip replacement. Instr Course
Lect 1995;44:293–304.
2. Schütz M, Perka C. Periprosthetic Fracture Management. Stuttgart and Thieme
New York: Thieme, 2013.
3. Gautier E, Sommer C. Guidelines for the clinical application of the LCP. Injury
2003;34(Suppl 2):B63-B76.
4. Dehghan N, McKee MD, Nauth A, Ristevski B, Schemitsch EH. Surgical fix-
ation of Vancouver type B1 periprosthetic femur fractures: a systematic review. J
Orthop Trauma 2014;28:721–727.
5. Moore RE, Baldwin K, Austin MS, Mehta S. A systematic review of open reduc-
tion and internal fixation of periprosthetic femur fractures with or without allograft
strut, cerclage, and locked plates. J Arthroplasty 2014;29:872–876.
6. Stoffel K, Sommer C, Kalampoki V, Blumenthal A, Joeris A. The influence of
the operation technique and implant used in the treatment of periprosthetic hip and
interprosthetic femur fractures: a systematic literature review of 1571 cases. Arch
Orthop Trauma Surg 2016;136:553–561.
7. Pavlou G, Panteliadis P, Macdonald D, et al. A review of 202 periprosthetic
fractures--stem revision and allograft improves outcome for type B fractures. Hip Int
2011;21:21–29.
8. Abdel MP, Cottino U, Mabry TM. Management of periprosthetic femoral fractures
following total hip arthroplasty: a review. Int Orthop 2015;39:2005–2010.
9. Füchtmeier B, Galler M, Müller F. Mid-term results of 121 periprosthetic femoral
fractures: increased failure and mortality within but not after one postoperative year.
J Arthroplasty 2015;30:669–674.
10. No authors listed. The Swedish Hip Arthroplasty Register. https://shpr.registercen-
trum.se/shar-in-english/the-swedish-hip-arthroplasty-register/p/ryouZwaoe (date
last accessed 20 August 2019).
11. Chatziagorou G, Lindahl H, Garellick G, Kärrholm J. Incidence and demograph-
ics of 1751 surgically treated periprosthetic femoral fractures around a primary hip
prosthesis. Hip Int 2019;29:282–288.
12. No authors listed. Klassifikation av vårdåtgärder (KVÅ) https://www.social-
styrelsen.se/utveckla-verksamhet/gemensam-informationsstruktur/klassificer-
ing-och-koder/kva/ (date last accessed 20 August 2019).
13. No authors listed. NHS Digital: data and information. https://digital.nhs.uk/
data-and-information (date last accessed 20 August 2019).
14. No authors listed. The Swedish Knee Arthroplasty Register. http://www.myknee.
se/en/ (date last accessed 20 August 2019).
15. Fink B, Oremek D. Hip revision arthroplasty for failed osteosynthesis in peripros-
thetic Vancouver type B1 fractures using a cementless, modular, tapered revision
stem. Bone Joint J 2017;99-B(4 Suppl B):11–16.
16. Abdel MP, Watts CD, Houdek MT, Lewallen DG, Berry DJ. Epidemiology of
periprosthetic fracture of the femur in 32 644 primary total hip arthroplasties: a
40-year experience. Bone Joint J 2016;98-B:461–467.
17. Lindahl H, Malchau H, Herberts P, Garellick G. Periprosthetic femoral fractures
classification and demographics of 1049 periprosthetic femoral fractures from the
Swedish National Hip Arthroplasty Register. J Arthroplasty 2005;20:857–865.
18. Witte D, Klimm M, Parsch D, et al. Ten-year survival of the cemented MS-30 femo-
ral stem: increased revision rate in male patients. Acta Orthop Belg 2009;75:767–775.
19. Hailer NP, Garellick G, Kärrholm J. Uncemented and cemented primary total hip
arthroplasty in the Swedish Hip Arthroplasty Register. Acta Orthop 2010;81:34–41.
20. Chatziagorou G, Lindahl H, Kärrholm J. The design of the cemented stem influ-
ences the risk of Vancouver type B fractures, but not of type C: an analysis of 82,837
Lubinus SPII and Exeter Polished stems. Acta Orthop 2019;90:135–142.
21. Konstantinidis L, Schmidt B, Bernstein A, et al. Plate fixation of peripros-
thetic femur fractures: what happens to the cement mantle? Proc Inst Mech Eng H
2017;231:138–142.
22. Springer BD, Berry DJ, Lewallen DG. Treatment of periprosthetic femoral frac-
tures following total hip arthroplasty with femoral component revision. J Bone Joint
Surg [Am] 2003;85-A:2156–2162.
23. Haddad FS, Duncan CP, Berry DJ, et al. Periprosthetic femoral fractures around
well-fixed implants: use of cortical onlay allografts with or without a plate. J Bone
Joint Surg [Am] 2002;84-A:945–950.
24. Moreta J, Uriarte I, Ormaza A, et al. Outcomes of Vancouver B2 and B3 peri-
prosthetic femoral fractures after total hip arthroplasty in elderly patients. Hip Int
2019;29:184–190.
25. Khan T, Grindlay D, Ollivere BJ, et al. A systematic review of Vancouver B2 and
B3 periprosthetic femoral fractures. Bone Joint J 2017;99-B(4 Suppl B):17–25.
26. Lindahl H, Garellick G, Regnér H, Herberts P, Malchau H. Three hundred and twen-
ty-one periprosthetic femoral fractures. J Bone Joint Surg [Am] 2006;88-A:1215–1222.
1458 G. Chatziagorou, H. Lindahl, J. Kärrholm 
27. Corten K, Vanrykel F, Bellemans J, et al. An algorithm for the surgical treatment
of periprosthetic fractures of the femur around a well-fixed femoral component. J
Bone Joint Surg [Br] 2009;91-B:1424–1430.
28. Brady OH, Garbuz DS, Masri BA, Duncan CP. The reliability and validity of the
Vancouver classification of femoral fractures after hip replacement. J Arthroplasty
2000;15:59–62.
Author information:
G. Chatziagorou, MD, Orthopaedic Surgeon, The Swedish Hip Arthroplasty
Register, Gothenburg, Sweden; Department of Orthopaedics, Institute of Clin-
ical Sciences, Sahlgrenska Academy, University of Gothenburg, Gothenburg,
Sweden.
H. Lindahl, MD, PhD, Consultant Orthopaedic Surgeon, The Swedish Hip
­Arthroplasty Register, Department of Orthopaedics, Lidköping Hospital, Lid-
köping, Sweden.
J. Kärrholm, MD, PhD, Professor, Consultant Orthopaedic Surgeon, The
Swedish Hip Arthroplasty Register, Department of Orthopaedics, Institute
of Clinical Sciences, Sahlgrenska Academy, University of Gothenburg,
Gothenburg, Sweden.
Author contributions:
G. Chatziagorou: Conceived and designed the study, Collected the data,
Performed the analysis, Wrote and edited the manuscript.
H. Lindahl: Conceived the study, Collected the data, Edited the manuscript.
J. Kärrholm: Conceived and designed the study, Performed the analysis, Edited
the manuscript.
Follow us @BoneJointJ
29. Lenz M, Perren SM, Gueorguiev B, et al. A biomechanical study on proximal
plate fixation techniques in periprosthetic femur fractures. Injury 2014;45(Suppl
1):S71-S75.
30. Lenz M, Windolf M, Mückley T, et al. The locking attachment plate for proximal
fixation of periprosthetic femur fractures--a biomechanical comparison of two tech-
niques. Int Orthop 2012;36:1915–1921.
Funding statement:
The first data-linking was supported financially by the Healthcare Subcommit-
tee, Region Västra Götaland, Sweden. The second crossmatching with the
National Patient Register (NPR) was supported by the Swedish Hip Arthroplasty
Register (SHAR), as was the linkage with the Swedish Knee Arthroplasty
Register (SKAR).
No benefits in any form have been received or will be received from a commer-
cial party related directly or indirectly to the subject of this article.
Acknowledgements:
The authors would like to acknowledge Erik Büllow for his contribution to
the statistical analysis of this study, and Otto Robertsson for assistance with
cross-referencing with the Swedish Knee Arthroplasty Register.
Ethical review statement:
The study was approved by the Central Ethical Review Board in Gothenburg
(entry number: 198-12; date: 5 April 2012).
This article was primary edited by D. Johnstone.
THE BONE & JOINT JOURNAL