ArticlePDF Available

Abstract and Figures

Introduction The uncemented total hip arthroplasty relies on a secure initial fixation of the femoral stem to achieve osseointegration. Undersizing of the femoral implant compromises this. Surgeons routinely review postoperative radiographs to assess appropriate sizing, but existing methods of assessment lack standardisation. We present a system of accurately and reliably classifying radiological undersizing, which will help us better understand the factors that might have led to undersizing. Aim To describe and evaluate a classification system for assessing radiological undersizing of the uncemented stem in hip arthroplasty. Method We conducted a retrospective review of 1,337 consecutive hip arthroplasties using the Corail stem. Two independent investigators reviewed post-operative radiographs and classified them as either appropriately sized or undersized. Undersized stems were sub-categorised into four subtypes: uniformly undersized, varus undersized, valgus undersized or ‘cocktail-glass’ undersized. Inter- and intra-observer agreement was determined. The accuracy of our classification system was validated by comparison with digital re-templating. We further assessed patient demographics and stem size in relation to sizing. Results 1 in 5 cases (19.75%) were deemed radiologically undersized. The commonest subtypes of undersizing were uniformly (47%) and varus (39%) undersized. When assessing sizing and subtype categorisation, inter-observer agreement was 89–92% and intra-observer agreement 86%. Classification decisions showed 92% and 97% accuracy for uniformly undersizing and varus undersizing respectively when validated against digital re-templating. Age, gender and smaller stem size were significantly associated with radiological undersizing. The Corail KLA model (125° neck) was found to have a higher incidence of stems undersized in varus. Conclusions This study describes and validates a classification system for the analysis of radiological undersizing.
Content may be subject to copyright.
HIP International
1 –10
© The Author(s) 2021
Article reuse guidelines:
DOI: 10.1177/1120700021996001
Uncemented stems are commonly used in both primary
and revision total hip arthroplasty (THA).1 Titanium
hydroxyapatite (HA) coated stems have been consistently
shown to achieve long term biological fixation.2–4 Studies
involving porous coated stems concluded that micromo-
tion <40 microns leads to osseointegration whereas
motion >150 microns leads to fibrous fixation.5,6 Animal
studies have shown HA coating more likely to achieve
Definition and validation of a system for
classifying the undersized Corail femoral
stem in total hip arthroplasty
Jamie S McConnell1, Farhan A Syed2, Paul Saunders2,
Raviprasad Kattimani3, Anthony Ugwuoke4, Merzesh Magra5
and Steve K Young2
Introduction: The uncemented total hip arthroplasty relies on a secure initial fixation of the femoral stem to achieve
osseointegration. Undersizing of the femoral implant compromises this. Surgeons routinely review postoperative
radiographs to assess appropriate sizing, but existing methods of assessment lack standardisation. We present a system
of accurately and reliably classifying radiological undersizing, which will help us better understand the factors that might
have led to undersizing.
Aim: To describe and evaluate a classification system for assessing radiological undersizing of the uncemented stem in
hip arthroplasty.
Method: We conducted a retrospective review of 1,337 consecutive hip arthroplasties using the Corail stem. Two
independent investigators reviewed post-operative radiographs and classified them as either appropriately sized or
undersized. Undersized stems were sub-categorised into four subtypes: uniformly undersized, varus undersized, valgus
undersized or ‘cocktail-glass’ undersized. Inter- and intra-observer agreement was determined. The accuracy of our
classification system was validated by comparison with digital re-templating. We further assessed patient demographics
and stem size in relation to sizing.
Results: 1 in 5 cases (19.75%) were deemed radiologically undersized. The commonest subtypes of undersizing were
uniformly (47%) and varus (39%) undersized. When assessing sizing and subtype categorisation, inter-observer agreement
was 89–92% and intra-observer agreement 86%. Classification decisions showed 92% and 97% accuracy for uniformly
undersizing and varus undersizing respectively when validated against digital re-templating. Age, gender and smaller stem
size were significantly associated with radiological undersizing. The Corail KLA model (125° neck) was found to have a
higher incidence of stems undersized in varus.
Conclusions: This study describes and validates a classification system for the analysis of radiological undersizing.
Hydroxyapatite, uncemented, undersizing
Date received: 19 September 2020; accepted: 21 December 2020
1 Barnet and Chase Farm Hospitals NHS Trust, London, UK
2 Warwick Hospital, Warwick, UK
3 Macclesfield District General Hospital, Macclesfield, UK
4 Countess of Chester Hospital, UK
5 University Hospitals of Morecambe Bay NHS Foundation Trust,
Kendal, UK
Corresponding author:
Farhan A Syed, South Warwickshire NHS Foundation Trust, Lakin
Road, Warwick, CV34 5BW, UK.
996001HPI0010.1177/1120700021996001HIP InternationalMcConnell et al.
Original Research Article
2 HIP International 00(0)
osteointegration when compared to porous coated stems in
presence of micromotion of 150 microns.7 Nevertheless,
the aim is to achieve primary stability (the interface fit) at
the time of the surgery.
Adequacy of stem size is assessed on postoperative
radiograph. Typically, this assessment is made on an anter-
oposterior (AP) radiograph. The term “undersized”
describes an observation where the femoral canal could
have accommodated a larger stem. Achievement of correct
sizing is influenced by the morphology of the femur, the
quality of the intramedullary bone and the surgical tech-
nique used.
Radiological undersizing is not synonymous with stem
instability, which is a clinical finding leading to excessive
micro motion. Radiological undersizing of the Corail stem
has been shown to be associated with the development of
radiolucent lines (RLL) and ultimately aseptic loosen-
ing.8,9 The most recent of these studies identified undersiz-
ing by more than two sizes as the only predictive factor for
the development of RLL in zone 7.8
Joint registry studies have shown that smaller unce-
mented stem sizes are at an increased risk of failure, how-
ever they did not determine whether those stems were
radiologically undersized.10,11 The clinical relevance of
radiological undersizing merits further investigation, how-
ever to do so, a consistent and validated system of classify-
ing radiological undersizing is required.
Current methods of classifying radiological undersizing
include comparisons of implanted stem size against preop-
erative and postoperative radiograph templating,8,9,12 or a
canal-fill index ratio that requires measurements at various
points along the stem.13,14 Both methods are time consum-
ing and rely on calibration of radiograph, thus making
them cumbersome to employ.
The primary purpose of this study was to describe a sys-
tem for quantifying radiological undersizing of an unce-
mented femoral stem (Corail; DePuy-Synthes, Warsaw,
IN, USA). The primary outcomes are the inter- and intra-
observer reliability of the classification system along with
an evaluation of the accuracy of the classification system
against re-templating. The secondary outcome is to exam-
ine the ways in which the stems were undersized.
The initial step was to identify the modes of undersizing,
i.e. the various radiographic appearances that could be
seen in stems that were undersized. To achieve this, the 2
lead investigators (SKY and JSM) reviewed 100 AP radio-
graphs of postoperative THAs, identifying the common
patterns of undersized Corail stems.
The Corail stem is available in 6 types with different
extramedullary geometry; 4 of which were used in this
cohort (Table 1). The stem relies on compaction broaching
for primary fixation. The operative technique states that
there should be a 1-mm margin of cancellous bone around
the implant.15 In practice this is an unworkable definition,
as not only will the contour of the patient’s bone rarely be
an exact match to the implant stem but the Corail implant’s
dimensions do not scale linearly with increasing size
(Table 2). For example, if the surgeon increased from a
size 8 to a size 9 stem, the width of the stem at the collar
and lateral point would increase by half a millimetre, how-
ever the mid-point would remain the same and the tip
width would decrease by nearly 1 mm, resulting in failure
to achieve the desired 1-mm margin.
In assessing the adequacy of femoral stem size, the
surgeon must consider the differences in stem geometry,
the variability in morphology (e.g. Dorr Classification)
and the inevitable minor variations in positioning of the
stem at the time of surgery. The classification system
was intentionally designed to be used with non-cali-
brated images. We thus identified subtypes of radiologi-
cal undersizing based on no prior knowledge of the
prosthesis (e.g. model, size), femur morphology type, or
image calibration of.
Definition of the 4 subtypes of undersizing requires
description of four specific landmarks of the Corail stem
(Figure 1). These are the:
Collar: present on KA and KLA stems only;
Lateral-point: proximal lateral point of the stem;
Mid-point: halfway along the long axis of the pros-
thesis, measured from the shoulder to the tip;
Table 1. Corail stem models and their anatomical variants.
Corail stem models Neck-shaft
Collar Off-set Neck
KA (Standard) 135° Yes 40.3 mm 39 mm
KS (Standard) 135° No 40.3 mm 39 mm
KLA (Coxa Vara) 125° Yes 47.1 mm 40 mm
KHO* (High Offset) 135° No 47.2 mm 43 mm
*The KHO stem, although now available with a collar, was not available
with a collar during the time period of the cases in this study.
Table 2. Dimensions of the Corail standard (KA) stem.
stem size
Length of
stem (mm)
Stem width (mm)
Collar Mid-
8 115 24.7 11.9 17.7 6.3
9 130 25.3 11.9 18.8 5.6
10 140 26.0 12.7 20.2 6.0
11 145 27.4 14.0 20.5 7.1
12 150 28.9 15.2 22.1 8.2
13 155 29.9 16.1 23.3 9.0
14 160 30.9 17.1 24.6 9.8
McConnell et al. 3
Tip: the point ‘one stem width’ from the distal end
of the prosthesis. This landmark is easy to visualise,
however for constancy a formal definition is: draw
a square as wide as the tip of the stem, with the bot-
tom aligned to the lower margin of the prosthesis,
the landmark point is where the top of the square
meets the axis of the prosthesis.
The classification system
The 4 patterns of radiological undersizing are summarised
as follows:
Uniformly undersized:
At no point there is contact of the stem against the cor-
tex. There is a margin of excess cancellous bone on both
medial and lateral aspects of the entire length of the stem.
In the majority of cases this is symmetrical, such that a
half-tip width will pass all around the stem. Where the
margin is not symmetrical, the combined medial and lat-
eral measurement will be equal to or greater than the width
of the stem tip (Figure 2).
4. Tip
2. Lateral-point
3. Mid-point
Figure 1. Landmark points on the Corail stem (DePuy,
Warsaw, IN, USA; with permission).
Figure 2. Uniformly undersized: (a) half tip width of cancellous bone around stem, (b) half tip width (tip landmark point described
in methods.
4 HIP International 00(0)
Varus undersized:
The stem is in varus with respect to the anatomical axis
of the femur. Proximally the stem is in contact with the
medial calcar and distally with the lateral cortex. This cre-
ates two areas of excessive cancellous bone around the
implant (Figure 3(a)): proximally between the stem and
the lateral cortex and distally between the stem and the
medial cortex. A triangle can be constructed in the distal
cancellous space, with the base of the triangle level with
the stem tip. This triangle is then copied and inverted into
the proximal lateral cancellous space, as distally as possi-
ble; the limiting factor will be the base of the triangle
against the lateral cortex and lateral aspect of the stem.
This can be done on non-calibrated images. The triangular
wedges must overlap by more than half the height of the
triangles to infer a meaningful undersizing (Figure 3(b)). If
the triangles do not overlap by half their height, then the
stem is not undersized (Figure 3(a)).
Valgus undersized:
This follows the same principles of the varus under-
sized definition. The stem is in contact proximally with
the lateral cortex, and distally with the medial cortex. A
triangle of excess cancellous bone is drawn in the distal
lateral space. The triangle is inverted into the proximal
medial space. If the triangles overlap by at least half their
height, then the stem is considered undersized (Figure
4(b)). If the triangles do not overlap by at least half their
height, the stem is not undersized (Figure 4(a)).
Cocktail-glass undersized:
The distal segment of the stem is well-sized within the
diaphysis yet has insufficient bony contact at the metaphy-
sis. Distally, the stem is in contact with the cortex.
Proximally, (at the level of the collar, shoulder point and
the midpoint), the stem has a margin of cancellous bone,
than a half-tip width at both medial and lateral aspects
(Figure 5).
We conducted a retrospective radiographic review of 1337
consecutive cases of uncemented primary THAs. All sur-
geries were performed by a Consultant grade fellowship-
trained arthroplasty surgeons (n = 4) between January 2009
and December 2013, using an uncemented titanium, HA
Figure 3. Varus stems (a) not undersized as no triangle overlap, (b) undersized due to >50% overlap of triangles.
McConnell et al. 5
coated femoral stem (Corail, De Puy Synthes, Warsaw,
IN). There were 523 males and 814 females. The mean age
of the patients was 70 years (range 24–94 years). The mean
body mass index (BMI) of was 29.1 kg/m2 (range 15–
60 kg/m2).
Two investigators independently reviewed the postop-
erative AP radiographs and classified each case according
to the classification system described above; undersized
cases were further classified according to the four sub-
types. A standardised protocol for radiographic assessment
was used: AP pelvis projection taken with feet internally
rotated, beam at 100-cm distance and centred at the mid-
point of the pubic symphysis. The first 350 cases were
reviewed twice by each reviewer at different time periods
in order to calculate both inter and intra-observer agree-
ment using Cohen’s Kappa co-efficient.16
Statistical analysis of differences in patients’ age, BMI
and stem sizes was conducted between appropriately sized
and undersized cases using the Mann-Whitney U-test.
Analysis of differences in age, BMI and stems sizes
between undersizing subtypes was performed using one-
way ANOVA. Chi-square analysis was conducted for
gender on undersizing and subtypes. Statistical analysis
was performed using SPSS Statistics for Windows soft-
ware (SPSS Inc., Chicago, Ill, USA) with a significance
level of p 0.05.
To determine sensitivity and specificity for uniformly
and varus undersizing, we re-analysed 200 radiographs
where the stem had been positioned either centrally or in a
degree of varus; i.e. not necessarily undersized according
to the classification system. Digital templating (TraumaCad
Digital Templating, Brainlab Inc, Ill) of the postoperative
radiograph was performed by a blinded assessor who was
unaware of the classification system, to determine the stem
size that would produce a 1-mm mantle. Stems were con-
sidered undersized on digital templating if the stem size
selected was 2 or more sizes greater than the size of stem
that had been implanted.
For uniformly undersized cases the cancellous bone
mantle was digitally measured at all points around the
stem and recorded as greater or less than the width of the
tip of the stem. For varus undersized cases, triangles were
digitally drawn, and the degree of overlap recorded as
greater or less than 50%.
Figure 4. Valgus stems (a) not undersized as no triangle overlap, (b) undersized due to >50% overlap of triangles.
6 HIP International 00(0)
Of the 1337 postoperative AP radiographs classified 264
stems were undersized. Of the undersized cases, 124 cases
were uniformly undersized, 103 cases were varus under-
sized, 19 cases were in valgus and 18 cases were catego-
rised as cocktail-glass undersized. The distribution of
sizing is presented in Figure 6.
Observer agreement
Analysis of classification agreement showed very good
inter-observer and good intra-observer agreement for both
undersizing and the sub-type categorisation of undersizing
(Table 3).
Comparison to re-templating
Digital templating of 200 cases was conducted to assess
the accuracy of the classification system for varus and
uniformly undersized; these 2 subtypes encompass 86% of
all undersized cases.
For the evaluation of potentially uniformly undersized
cases, in 92% where there was a transverse cancellous mar-
gin greater than the width of the tip, the stem inserted was
indeed two or more sizes undersized compared to the re-
templated image (true positive; uniformly undersized). In
contrast, in only 5% of cases where the cancellous space
was less than a tip width was the stem undersized by at least
2 sizes (false negative). Chi-square analysis demonstrates a
significant relationship between the cancellous space
(greater than the width of the tip of the stem) and whether
undersized by 2 or more on re-templating (p = < 0.0001).
The ‘uniformly undersized’ classification was calculated to
have a sensitivity of 92% and specificity of 95% (Table 4).
For the evaluation of potentially varus undersized
cases, in 97% of cases where triangle overlap was than
50% were cases undersized by two or more sizes on re-
templating (true positive; varus undersized). In contrast,
when there was <50% overlap only 9% were undersized
Figure 5. Cocktail glass (a) not undersized as a half tip width of cancellous bone is not present medially, (b) undersized as
proximally there is both medial and lateral free cancellous bone greater than a half tips width.
McConnell et al. 7
by two or more (false positive). Chi-square analysis dem-
onstrates a significant relationship between triangle over-
lap (50%) and undersizing by 2 or more on templating
(p < 0.00001). The ‘varus undersized’ classification was
calculated to have a sensitivity of 83% and specificity of
91% (Table 5).
The effect of patient and prosthesis factors on
The mean age of patients having an undersized stem was
significantly lower than the mean age of patients with
appropriately sized stems (68 vs. 70 years; p = 0.024, effect
size = 0.06) (Table 6). BMI did not differ between under-
sized and appropriately sized cases (p = 0.66). Undersizing
was more common in males (p = <0.00001). The mean
stem size in undersized cases was significantly lower than
the mean stem size for appropriately sized cases
(p = <0.000001; effect size = 0.31). Over half of all size 8
(56%) and 9 (51%) stems were deemed to be undersized. In
contrast, only 19% of all stems greater than a size 9 were
classed as undersized (Table 7).
With regard to differences in the sub-types of undersiz-
ing, there was a trend for uniformly and cocktail-glass
undersizing relatively more frequently in younger patients,
however no statistically significant difference was found for
undersizing sub-type and age (p = 0.17; f-ratio = 1.69). Mean
BMI was not significantly different across the subtypes
(p = 0.55; f-ratio = 0.70). The lowest mean BMI was in val-
gus cases, however this subtype comprised less than 2% of
the cohort (n = 19 only). There was a significant relationship
between gender and type of undersizing, with male patients
having the greatest number of varus-undersized and cock-
tail-glass undersized stems (Chi-square statistic = 46.58;
p = <0.00001). The mean stem size was significantly differ-
ent between undersizing sub-types (p = 0.012; f-ratio = 3.74),
with the lowest mean stem size in uniformly undersized
cases and the highest mean in cocktail-glass undersizing.
Stem type
The majority of stems in this cohort were the standard KA
(69% of cohort). Table 1 defines the various stem types;
Table 8 shows the distribution of undersizing relative to
stem type. Only 3% of cases involved KHO stems, and just
14% of stems were uncollared. A relatively high percent-
age of the lateralised (KLA) stems were undersized
Appropriately Sized
Uniformly Undersized
Varus Undersized
Valgus Undersized
Champagne Fluted Undersized
Figure 6. Classification of undersizing and mix of undersizing subtypes.
Table 3. Observer agreement for undersizing classification
Agreement of
Agreement of sub-
type of undersizing
92% 89%
Inter-observer Kappa Very good (0.826) Very good (0.818)
86% 86%
Intra-observer Kappa Good (0.675) Good (0.767)
Table 4. Digital templating compared to uniformly undersized
undersized < 2
undersized 2
Not Undersized (cancellous
space less than tip width)
61 3
Uniformly Undersized
(cancellous space greater
than tip width)
3 33
Table 5. Digital templating compared to varus undersized
undersized < 2
undersized 2
Not Undersized (triangle
overlap < 50%)
64 6
Varus Undersized (triangle
overlap > 50%)
1 29
8 HIP International 00(0)
compared to other stem models. The KLA stem (125° neck
shaft angle) was most likely to be undersized in varus. The
KHO was the least likely to be undersized. The standard
stem type (KA) was most likely to be uniformly under-
sized, as were the uncollared standard stem (KS) and the
high offset stem (KHO).
We have proposed a classification system that is reproduc-
ible, demonstrates good to very good observer agreement,
has high sensitivity and specificity, and can be applied
without the need for digital measurement. Defining a clas-
sification system for undersizing is the first step towards
enabling the further study of the relationship between radi-
ological assessment of sizing and long-term outcomes.
Further study will be needed to determine whether the
subtypes have any clinical relevance. We hypothesise that
since valgus or varus undersized stems have contact with
the cortex, these subtypes will experience less micromo-
tion and thus will be less likely to demonstrate subsidence
or aseptic loosening on future imaging. Uniformly under-
sized stems would theoretically be of greatest risk of
loosening due to excessive micromotion and consequently
poor osseointegration. In our series, uniformly undersized
cases had a trend to be older, female and have smaller stem
sizes inserted. Evidence from national registries has con-
sistently shown that smaller uncemented stems have a
greater risk for revision than larger sized stems.11,17 The
relative revision rate, calculated from over 40,000 unce-
mented hydroxyapatite coated stem cases, was 7.7% for
size 8 or 9 stems, compared to just 3% for stem sizes 10–
20. Our data shows a trend for greater radiological under-
sizing in smaller stem sizes (Figure 7).
Another important subtype to highlight is the cocktail-
glass undersizing. These stems have a metaphyseal - dia-
physeal mismatch, in that they appear undersized proximally
but well fixed, or even over-sized, distally. A recent single
centre review of 53 size 8/9 Corail stems found no increased
failure rate in smaller stems, however did highlight that
smaller stem sizes had a tendency to be used in younger
patients.18 The authors suggest that the observed failure rate
in smaller stem sizes seen in registry studies may be due to
altered anatomy associated with younger patients e.g,
thicker cortices and ‘Dorr A’ femur type.18 Recently, the
importance of canal diameter as opposed to the Dorr
Table 6. The effect of patient and prosthesis factors on undersizing.
Factors Appropriately
Undersized Uniformly
Cases (total) 1073 264 124 103 19 18
Mean age (years) 70 68 67 70 69 66
Mean BMI (kg/m2) 28.9 29.5 29.7 29.4 26.8 30.3
Female (%) 64 48 64 27 84 17
Mean stem size 11.9 10.6 10.4 10.9 10.5 11.5
BMI, body mass index.
Table 7. Undersizing of different sized stems.
Stem sizes 8 9 10 11 12 13 14 15 16 18
Total 43 74 184 362 310 185 90 58 25 6
Appropriately sized 19 36 114 278 276 165 76 51 25 6
Undersized 24 38 70 84 34 20 14 7 0 0
% Undersized 56 51 38 23 11 11 16 12 0 0
Table 8. Radiological undersizing by Corail stem type.
Stem type Cases Undersized
KA (C) 927 155 (17%) 82 (53%) 50 (32%) 12 (8%) 11 (7%)
KS 136 33 (24%) 20 (61%) 5 (15%) 4 (12%) 4 (12%)
KLA (C) 230 71 (31%) 19 (27%) 47 (66%) 2 (3%) 3 (4%)
KHO 44 5 (11%) 3 (60%) 1 (20%) 1 (20%) 0 (0%)
(C), collared stems.
McConnell et al. 9
classification has been described.19 Further investigation
into the association between femur morphology and/or
canal diameter with clinical outcomes is warranted.
We hypothesise that for the reasons discussed above,
the different subtypes of undersizing may lead to a differ-
ent presentation of radiological and clinical complications
on follow-up. Ultimately, if it is determined that undersiz-
ing, or particular subtypes of undersizing, predispose to
increased failure risk, this will need to be taken into
account in patient selection, prosthesis selection and pre-
operative templating parameters that may be able to eradi-
cate this observed effect.
Templating studies have reported on undersizing
when the degree is by 2 or more sizes.9 We have demon-
strated that visual assessment alone can consistently
identify undersizing with near equivalent accuracy to
digital re-templating. When cases conformed to the clas-
sification definitions, namely the presence of overlap-
ping triangles (>50%) in varus positioned stems and a
continuous tip width cancellous space in stems posi-
tioned centrally, accurate identification of an implant
two sizes too small (on re-templating) was demonstrated
in 92–97% of cases.
A significant relationship between age and gender in
the presence of radiological undersizing was observed. No
relationship was identified between BMI and radiological
undersizing. These findings suggest that there might be
reasons other than radiological undersizing that has lead to
increased failure in patients with BMI >30 kg/m2 , as
reported by Jameson et al.17
The lateralised high-offset stem (KLA) had a higher
prevalence of radiological undersizing. Within this stem
type, the highest proportion of undersized cases were for
the varus subtype. Implanting a high offset stem into rela-
tive varus can drastically increase the femoral offset.
Although not within the remit of this paper, we hypothe-
sise that this is related to the observed finding that lateral-
ised stems have a higher rate of revision secondary to
aseptic loosening in uncemented stems.20–22
A limitation of this study is that our classification
method does not consider the appearance on the lateral
radiograph. This decision was taken as in everyday prac-
tice, we commonly see that only AP radiographs are per-
formed at post-operative assessments. We found lateral
radiographs to vary in quality as it can be difficult to
obtain good-quality lateral radiographs in the acute post-
operative phase, as positioning can be restricted by pain
and stiffness. No cases were excluded in this cohort study
because of poor AP positioning on any radiograph in order
to eliminate the risk of experimenter bias. This paper does
not consider the clinical implications of undersizing; cre-
ating a validated classification system is the first step in
answering this clinical question. In relation to clinical out-
comes, a potential limitation with the present study is it
that it does not gauge the degree of undersizing, i.e. we
did not quantify how many stem sizes smaller our under-
sized stems were, compared to the ideal size. There may
be an incremental effect of undersizing on clinical out-
comes, i.e. the greater the degree of undersizing the
greater the complication rate.
In conclusion we have described and validated a clas-
sification system for radiological undersizing with an
uncemented femoral stem. This is the first step in consider-
ing the question of whether radiological undersizing of an
uncemented femoral stem results in future radiological
and/or clinical problems.
020406080 100
Stem Size
% Undersized % Appropriately Sized
Figure 7. Percentage of radiological undersizing by stem size.
10 HIP International 00(0)
Declaration of conflicting interests
The author(s) declared the following potential conflicts of inter-
est with respect to the research, authorship, and/or publication of
this article: SKY: reports personal fees and non-financial support
from Depuy, outside the submitted work.
All other authors declare that there is no conflict of interest.
The author(s) received no financial support for the research,
authorship and/or publication of this article.
1. National Joint Registry for England, Wales and Northern
Ireland. NJR 16th annual report, http://www.njrcentre.
Default.aspx (2019, accessed 29 September 2019).
2. Chambers B, St Clair SF and Froimson MI. Hydroxyapatite-
coated tapered cementless femoral components in total hip
arthroplasty. J Arthroplasty 2007; 22: 71–74.
3. Reikerås O. Total hip arthroplasty with a fully hydroxyapa-
tite-coated stem: a cohort study during 23–28 years. J
Arthroplasty 2017; 32: 1543–1546.
4. Jacquot L, Bonnin MP, Machenaud A, et al. Clinical and radi-
ographic outcomes at 25–30 years of a hip stem fully coated
with hydroxylapatite. J Arthroplasty 2018; 33: 482–490.
5. Engh CA, O’Connor D, Jasty M, et al. Quantification of
implant micromotion, strain shielding, and bone resorption
with porous-coated anatomic medullary locking femoral
prostheses. Clin Orthop Relat Res 1992; 285: 13–29.
6. Jasty M, Bragdon C, Burke D, et al. In vivo skeletal
responses to porous-surfaced implants subjected to small
induced motions. J Bone Joint Surg Am 1997; 79: 707–714.
7. Søballe K, Hansen E, Brockstedt-Rasmussen H, et al.
Hydroxyapatite coating converts fibrous tissue to bone
around loaded implants. J Bone Joint Surg Br 1993; 75:
8. Magill P, Hill J, O’Brien S, et al. Observed effect of femoral
component undersizing and a collarless design in the devel-
opment of radiolucent lines in cementless total hip arthro-
plasty. Arthroplast Today 2020; 6: 99–103.
9. Magill P, Blaney J, Hill JC, et al. Impact of a learning curve
on the survivorship of 4802 cementless total hip arthroplas-
ties. Bone Joint J 2016; 98–B: 1589–1596.
10. Jameson SS, Baker PN, Mason J, et al. Independent pre-
dictors of failure up to 7.5 years after 35 386 single-brand
cementless total hip replacements. Bone Joint J 2013; 95–B:
11. Hoskins WT, Bingham RJ, Lorimer M, et al. The effect
of size for a hydroxyapatite-coated cementless implant on
component revision in total hip arthroplasty: an analysis of
41,265 stems. J Arthroplasty 2020; 35: 1074–1078.
12. Rivera F, Leonardi F, Evangelista A, et al. Risk of stem
undersizing with direct anterior approach for total hip
arthroplasty. Hip Int 2016; 26: 249–253.
13. Aldinger PR, Jung AW, Breusch SJ, et al. Survival of
the cementless Spotorno stem in the second decade. Clin
Orthop Relat Res 2009; 467: 2297–2304.
14. Fottner A, Woiczinski M, Kistler M, et al. Influence of
undersized cementless hip stems on primary stability and
strain distribution. Arch Orthop Trauma Surg 2017; 137:
15. Vidalain JP, Selmi TA, Beverland D, et al. (eds.). The
CORAIL® Hip System: a practical approach based on 25
years of experience. Berlin, Heidelberg: Springer-Verlag,
16. Cohen J. A coefficient of agreement for nominal scales.
Educ Psychol Meas 1960; 20: 37–46.
17. Jameson SS, Baker PN, Mason J, et al. Independent pre-
dictors of failure up to 7.5 years after 35 386 single-brand
cementless total hip replacements: a retrospective cohort
study using National Joint Registry data. Bone Joint J 2013;
95–B: 747–757.
18. Piggott RP, Lyons R, Murphy CG, et al. Total hip replace-
ment revision in a single brand small cementless stem – our
experience after the findings of the National Joint Registry.
Arch Bone Jt Surg 2018; 6: 501–507.
19. Karayiannis PN, Cassidy RS, Hill JC, et al. The relation-
ship between canal diameter and the Dorr classification. J
Arthroplasty 2020; 35: 3204–3207.
20. Courtin C, Viste A, Subtil F, et al. Cementless lateralized
stems in primary THA: mid-term survival and risk factors
for failure in 172 stems. Orthop Traumatol Surg Res 2017;
103: 15–19.
21. Cantin O, Viste A, Desmarchelier R, et al. Compared fixa-
tion and survival of 280 lateralised vs 527 standard cement-
less stems after two years (1–7). Orthop Traumatol Surg
Res 2015; 101: 775–780.
22. Ramaniraka NA, Rakotomanana LR, Rubin PJ, et al.
[Noncemented total hip arthroplasty: influence of extramed-
ullary parameters on initial implant stability and on bone-
implant interface stresses]. Rev Chir Orthop Reparatrice
Appar Mot 2000; 86: 590–597.
... To assess stem-bone contact, a digital templating software (ZedHip, Lexi, Tokyo, Japan) was used to overlay pre-and post-operative CT images using reference points, setting the cortico-cancellous bone interface at 600 Hounsfield units (HU). Contact was defined as > 599HU, but small contact areas such as isolated points or lines Fig. 1 The Corail stem is available in a variety of sizes (6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19)(20) and in five geometries ( shorter than 5 mm were ignored. The digital templating software was not able to assess contact between the collar and cortical bone, thus this was assessed directly from postoperative CT scans. ...
... The following binary assessments were made (yes, 1; no, 0): contact between the collar and proximal cortical bone, contact between the stem and proximal cortical bone, contact between the stem and distal cortical bone. Stem-bone contact patterns were inspected around the proximal (Gruen zones 1, 2, 6,7,8,9,13,14) and distal (Gruen zones 3,4,5,10,11,12) portions of the stem, and classified a posteriori, as follows ( Fig. 6): type 1 was defined as neither proximal nor distal cortical contact, type 2 was defined as only proximal cortical contact, type 3 was defined as only distal cortical contact, type 4 was defined as one proximal and one distal cortical contact area, and type 5 was defined as at least one proximal cortical contact area and more than one distal cortical contact area. ...
... This suggests that sagittal and coronal morphology may influence stem fixation and positioning, though sagittal morphology is often ignored, and coronal morphology is predominantly used for surgical planning and templating. The means of obtaining initial fixation for a number of uncemented stem designs has been previously described [10], with several 2D radiographic studies evaluating the initial fixation of long straight tapered stems that use the compaction broaching technique [7,11,12]. One of these studies demonstrated that the intraoperative positioning of the stem depends on preoperative femoral morphology, [12] while another study showed that preoperative femoral bone morphology influences the risk of calcar fractures [7], and a final study found that 1 in 5 stems are undersized, although the effect of under-sizing on postoperative outcomes was not investigated [11]. ...
Full-text available
Introduction Despite excellent long-term survival, total hip arthroplasty (THA) can fail due to aseptic loosening, dislocations, sepsis and periprosthetic fractures, all of which remain considerably burdensome. Aseptic loosening is one of the main causes of THA failure, often due to osteolysis, stress shielding and/or lack of primary stability. This study aimed to investigate stem-bone contact patterns of a long straight-tapered uncemented stem following primary THA, and to determine whether these contact patterns are related to preoperative femoral morphology and whether they influence postoperative outcomes. Materials and methods The authors reviewed a continuous series of 60 hips (55 patients) that underwent primary THA using the Corail® stem (DePuy, Leeds, UK). Patients were evaluated pre- and post-operatively using the Japanese Orthopaedic Association (JOA) score. Computed-tomography (CT) scans were performed preoperatively to assess femoral bone morphology, and immediate postoperatively to assess stem-bone contact patterns. Postoperative radiographs were performed to calculate the Engh score. Regression analyses were performed to determine associations of postoperative JOA and Engh score with 27 independent variables. Results Forty-nine patients (54 hips) were assessed at 31 ± 8 months, with a JOA score of 92.9 ± 8.1 and an Engh score of 21.2 ± 1.9. Six patients (6 hips) were lost-to-follow-up. There were no revisions and only one complication (recurrent dislocation). Stem-bone contact patterns were associated with preoperative femoral morphology (sagittal CFI [p = 0.006], femoral offset [p = 0.028], and NSA [p = 0.022]), but were not associated with either postoperative JOA or postoperative Engh score. Conclusions The stem-bone contact patterns of a long straight-tapered uncemented stem are related to preoperative femoral morphology, but do not influence short-term postoperative outcomes. Contact patterns were related to preoperative femoral offset, NSA, and sagittal CFI, but not coronal CFI. Surgeons should, therefore, consider sagittal morphology for surgical planning and templating, in addition to the conventional parameters of coronal morphology.
... This has 95% coverage for the entire country. The radiographs were analyzed for stem alignment, Dorr classification on preoperative radiographs [7], and gross undersizing [8]. This corresponded to a uniformly contiguous margin around the stem. ...
Full-text available
Background To assess outcomes after uncemented hemiarthroplasty stems in the treatment of intracapsular femoral neck fractures over an 11-year period. Mortality rates were assessed, and whether proximal femoral geometry and stem alignment were factors in intraoperative or postoperative periprosthetic fracture (PPF) needs to be identified. Materials and Methods A retrospective single-center observational study was conducted of all patients who underwent intracapsular femoral neck fracture treatment using an uncemented prosthesis between January 2008 and December 2018. Primary endpoints included mortality rate, Dorr classification, prosthesis alignment, intraoperative fracture, and reoperation rate for any reason. Subanalysis on collared and uncollared implants was also conducted. Multivariate logistic regression was performed based on Dorr classification for stem alignment, fracture incidence, reoperation rate, implant sizing, and patient mortality. Results A total of 536 patients received an uncemented hemiarthroplasty in the study period. The mean patients age was 80.4 years, of which 71% were female. The 30-day mortality rate was 5.2%, with no deaths on day zero or day one. Twenty patients (3.7%) sustained an intraoperative calcar fracture, and 14 patients a PPF (2.6%) at a mean of 1.3 years from surgery. Dorr C type femurs were more likely to develop a PPF (P = .001), while valgus stem alignment was associated with PPF (P = .049). Conclusions This implant has low reoperation rates, low early postoperative mortality, and low 30-day mortality. This large single-center study provides up-to-date information using a contemporary stem in patients with multiple comorbidities. Dorr C femoral morphology and valgus stem malalignment were risk factors for postoperative fractures.
When using the anterior-based muscle sparing (ABMS) approach in total hip arthroplasty (THA), there are no limitations in the types of implants that can be utilized, including both cemented and cementless fixations, with no limitation in implant shape or design. Similar to all THA approaches, successful arthroplasty depends on accurate component positioning, achieved safely and efficiently so as to reconstruct appropriate hip joint mechanics. Further, the decreased risk of instability achieved with this approach allows routine usage of any size femoral head and aids in stability even for patients with abnormal spinopelvic motion. In this chapter I will discuss the reasoning behind why the ABMS approach allows universal implant application. Next, I offer an example of implants used at a successful ABMS program and discuss slight variations in technique that may be of interest to the novice as well as the experienced ABMS surgeon that aid in universal implant application.
Full-text available
Background The objective of this study was to determine the prevalence of radiolucent lines (RLLs) around the femoral component in a cohort of patients who underwent well-functioning cementless total hip arthroplasty (THA). Methods A cohort of unrevised Corail (DePuy Synthes, Raynham, MA) femoral components (n = 636) were analyzed at a median follow-up of 6.0 years (interquartile range: 5.2-6.8) with the Oxford Hip Score (OHS) and radiographs. Two independent observers assessed the radiographs for the presence of RLLs. Results The overall prevalence of RLLs in zone 7 was 13% (83/636). Patients with RLLs in zone 7 had an average OHS of 40.3 (15-48), and those who did not have RLLs in zone 7 had an average OHS of 38 (6-48), P = .07. Both groups had an average pain score of 1.6 out of 5, P = .5. The prevalence of RLLs in zone 7 was much less in the collared femoral components (2.6% prevalence) than in the collarless components (23.6% prevalence), but there was heterogeneity between these 2 groups preventing comparison. Logistic regression analysis of only the collarless components identified undersizing as the only predictive (odds ratio = 2.6) factor for RLL development in zone 7. Conclusions Undersizing the Corail stem is strongly predictive of developing RLLs in zone 7. Preoperative templating for the appropriate size is critical. We observed more RLLs in zone 7 with the collarless design Corail, but a comparison study with the same bearing couple is needed to investigate this further.
Full-text available
Introduction: Undersizing of cementless hip stems is a risk factor for aseptic loosening and early subsidence. The purpose of this study was to evaluate the effects of undersized stems and determine whether a biomechanical study can predict the clinical results. Materials and methods: Three consecutive sizes of a clinically proven stem (CLS Spotorno) were implanted into six composite femora (size large, Sawbones(®)), respectively. According to the Canal Fill Index (CFI), two stems (size 11.25 and 12.5) were undersized (CFI < 80%) and one stem (size 13.75) had an appropriate size (CFI > 80%). The primary stability was evaluated by measurement of 3-dimensional (3D)-micromotions under physiological adapted load and surface strains were recorded before and after implantation to detect stress-shielding processes. Results: Both undersized stems revealed significantly higher micromotions in all regions compared to the appropriate stem. The highest micromotions were registered at the distal tip of the three stem sizes. The changes in surface strain did not show a significant difference between the three stem sizes, but the highest strain reduction was observed proximally indicating a tendency for stress shielding. Conclusions: This study confirms the clinical assumption that undersized stem result in a significantly reduced primary stability. Furthermore, in vitro studies allow to determine the effects of undersizing and stress shielding processes.
Background Particularly in broach only uncemented total hip arthroplasty (THA) a narrow femoral canal presents a technical challenge. Traditionally such femurs have been considered to be Dorr A. To our knowledge, however, no study has reported on the relationship between isthmus width and the Dorr classification. Methods We reviewed 500 high quality hard copy radiographs. Dorr classification and isthmus canal width were measured using an electronic caliper by five independent observers with intra- and inter-observer error calculated. For this study we defined a narrow canal as being ≤10mm at its narrowest point (isthmus). Results Eight percent (40) were Dorr A, 85% (424) Dorr B and 7% (36) Dorr C. With respect to isthmus width for Dorr A 63% (25) were ≤10mm compared to just 13% (55) of Dorr B. However, overall because there were more Dorr B femurs, 69% of those with an isthmus of ≤10mm were Dorr B. Conclusion In this population almost 70% of patients with an isthmus ≤10mm were Dorr B, with only 30% being Dorr A. When using a broach only technique, isthmus width should be routinely measured on the pre-operative antero-posterior radiographs so as to alert the surgeon to potential problems.
Background: The cementless Corail is one of the most commonly used stems in total hip arthroplasty (THA). The aim of this study was to investigate whether there was a difference in revision rate for smaller stems. Methods: All primary THA procedures recorded by the Australian Joint Replacement Registry from September 1999 to December 2017 performed for osteoarthritis using the Corail stem, a cementless acetabular cup, modern bearing surfaces (ceramic/ceramic, ceramic/cross-linked polyethylene, and metal/cross-linked polyethylene), and 28 mm, 32 mm, and 36 mm head sizes were included. The primary outcome measure was femoral component revision. Data were analyzed and adjusted for age, gender, and head size. Further analysis investigated the effects of surgical approach. Results: There was 41,265 primary THAs recorded. The cumulative percent revision (CPR) at 13 years was 7.7% (5.5, 10.7) for stem sizes 8 and 9 and 3.0% (2.4, 3.8) for sizes 10-20 (P < .001). When adjusted for age and gender, the sizes 8 and 9 collared (hazard ratio [HR]: 6.22 [3.84-10.06], P < .001) and collarless (HR: 3.28 [2.41-4.45], P < .001) had a higher CPR than the collared and collarless size 10-20. The size 8 and 9 stems performed with an anterior approach had the highest CPR (HR: 14.44 [6.21-33.56], P < .001). The main reason for revision of size 8 and 9 femoral stems was loosening (65.2%, compared to 31.5% for 10-20 femoral stems). Conclusions: Smaller Corail stems have 4 times the rate of revision compared with the larger femoral sizes with loosening being the most common diagnosis. This is most evident when using an anterior approach.
Background: Cementless total hip replacement is the common THR performed in England, Wales, Northern Ireland and the Isle of Man. The Corail stem is the most popular cementless implant and has a ODEP 10A rating. Review of its performance in the registry identified an increase rate of revision amongst the smaller stem sizes. However, clarity was not provided on the explanation for this finding. We reviewed our own experience of smaller stems with a view to understanding the reasons for revision. Methods: We reviewed a single centre, single surgeon experience of the smaller Corail stem sizes for a ten-year period from 2003 to 2013. All data was collected from a prospectively maintained database. Details of clinical and radiological follow up were collected for all patients who had Corail stem size 8 and 9 implanted. Revision for any cause was taken as our endpoint. Results: 542 patients underwent total hip arthroplasty using the Corail stem during the study period. 53 small size Corail stems were implanted. The average age was 59 (range 17-88 years) and the average follow up was 41.4 months (range 1-118 months). 6 patients underwent revision during the study period, but only 4 stems required revision. The reasons for revision were aseptic loosening, fracture and metal-on metal complications. Only two stems required revision for stem related factors (3.8%). Conclusion: There was no evidence of an increased rate of revision in the small Corail stems in our cohort.
The Corail® Hip System, one of the first to use hydroxyapatite, was developed in 1986 as an innovative solution for hip arthroplasty. It has since evolved to become one of the most used hip systems across the world, with more than 110,000 patients benefiting from the treatment each year. This book is designed as a practical manual that will serve both as a reference for surgeons training on the system and as a source of information, tips, and tricks for the more experienced who wish to learn from the cases of other surgeons. The book is divided into three main parts. The first part discusses everything that is practical about the system, including the basic science, surgical technique, treatment of complications, and the results achieved in large cohorts of patients. A full discussion of revision arthroplasty using the Corail family of stems is also included. The second part is devoted to the important issues of surgical approach, bearing options, acetabular preparation and, cup orientation and fixation. The final part focuses on the management of the patient prior to and after surgery. Health economic implications of using the Corail Hip System are reviewed, and a collection of standard and complex clinical cases is provided to which surgeons can refer when planning surgery. The book closes by examining the next 25 years for the Corail® Hip System including the future in developing countries.
Background: Little is known about the survival of total hip arthroplasty implants with bioactive coatings beyond the first 20 years. The authors aimed to report survival of a tapered hip stem fully coated with hydroxylapatite (HA) at follow-up of 25-30 years. Methods: Of the original series of 320 patients (347 hips), 12 patients (12 hips) had stem and cup revisions, 54 patients (55 hips) had cup revisions, 17 patients (17 hips) had liner exchange. A total of 207 patients (225 hips) died with stems in place and 21 patients (24 hips) could not be reached. This left a cohort of 80 patients (86 hips) with their original stem for assessment. Survival was analyzed using the Kaplan-Meier (KM) method and cumulative incidence function (CIF). Results: Considering stem revision as endpoint, the revision risk calculated using the KM method was 6.3%, whereas using the CIF it was 3.7%. Considering any reoperation as endpoint, the revision risk calculated using the KM method was 41.2%, whereas using the CIF it was 25.9%. The Harris Hip Score for 77 patients (18 hips) was 81.6 ± 15.2. Standard x-rays were available for 52 hips (49 patients), and 10 (19.2%) showed radiolucencies <2 mm thick. Conclusion: This study is the first to report outcomes of an HA-coated stem beyond 25 years. The survival of stem compares favorably with long-term survival of the Charnley cemented stem, and with shorter-term registry studies. The stem achieved its intended purpose of total osteointegration in the long-term, although the proximolateral region remains susceptible to radiolucencies.
Background: Long-term results of a hydroxyapatite (HA)-coated stem are sparse. We have followed a cohort of patients operated with a grit-blasted titanium stem designed for press-fit insertion and entirely plasma-sprayed with HA up to 28 years. Methods: In the years 1988-1993, we performed 323 primary total hip arthroplasties in 276 patients (189 women). Their mean age was 48 (15-79) years. During the following years, 83 patients with 88 hips have died, and 18 patients did not attend the follow-up examination, but had no major symptoms according to telephone interviews and control examination. Thus, 212 patients (255 hips) were followed up for more than 20 years. Results: Three patients were revised because of late periprosthetic infection, 1 because of a fall with periprosthetic fracture, and 2 stems were revised due to mechanical failure. Osteolysis was significantly associated with wear, and wear was significantly associated with the size of the femoral head. We found a small amount of proximal bone loss and a low incidence of distal hypertrophy of the bone. The bone changes confirmed a well-fixed femoral component in asymptomatic patients. Conclusion: Our findings indicate an essentially physiological weight distribution from the stem to the femoral bone with no significant thigh pain. The changes in the bone confirmed that the femoral component was well-fixed, and we conclude that a fully HA-coated titanium stem designed for press-fit insertion lasts for 23-28 years.
Aims: Our aim was to report survivorship data and lessons learned with the Corail/Pinnacle cementless total hip arthroplasty (THA) system. Patients and methods: Between August 2005 and March 2015, a total of 4802 primary cementless Corail/Pinnacle THAs were performed in 4309 patients. In March 2016, we reviewed these hips from a prospectively maintained database. Results: A total of 80 hips (1.67%) have been revised which is equivalent to a cumulative risk of revision of 2.5% at ten years. The rate of revision was not significantly higher in patients aged ≥ 70 years (p = 0.93). The leading indications for revision were instability (n = 22, 0.46%), infection (n = 20, 0.42%), aseptic femoral loosening (n = 15, 0.31%) and femoral fracture (n = 6, 0.12%). There were changes in the surgical technique with respect to the Corail femoral component during the ten-year period involving a change to collared components and a trend towards larger size. These resulted in a decrease in the rate of iatrogenic femoral fracture and a decrease in the rate of aseptic loosening. Conclusion: The rate of revision in this series is comparable with the best performing THAs in registry data. Most revisions were not directly related to the implants. Despite extensive previous experience with cemented femoral components, the senior author noted a learning curve requiring increased focus on primary stability. The number of revisions related to the femoral component is reducing. Any new technology has a learning curve that may be independent of surgical experience. Cite this article: Bone Joint J 2016;98-B:1589-96.
Introduction: Increasing the femoral offset when performing total hip arthroplasty (THA) theoretically increases the stresses and risks of the stem not integrating itself into bone. But this concept has not been validated for cementless stems; this led us to conduct a retrospective study to determine: (1) the risk factors for the occurrence of symptomatic femoral radiological abnormalities, (2) the incidence of these abnormal radiological findings, (3) the revision rate for aseptic non-integration of a cementless lateralized stem. Hypothesis: Young patients with significant femoral canal flare and a small cementless lateralized stem have a higher risk of abnormal osseointegration. Material and methods: We analyzed retrospectively 172 consecutive lateralized stems (KHO, Corail™ product line) implanted during primary THA between 2006 and 2012 in 157 patients (mean age 68years±12.6 (20-95), 89% men). Radiographs were used to evaluate osseointegration scores, offset restoration and the Noble index. Kaplan-Meier survival analysis was performed using "symptomatic femoral radiological abnormalities" and "revision for aseptic stem non-integration" as endpoints. Results: The mean follow-up was 5.9years±2.7 (range, 2-12.4years). Being more than 70years of age (HR=0.7, 95% CI: [0.3-0.9], P=0.004) and having a larger stem (HR=0.6, 95% CI: [0.4-0.9], P=0.03) were protective against symptomatic femoral radiological abnormalities, while increasing the postoperative femoral offset (HR=1.1, 95% CI: [1.01-1.2], P=0.02) was deleterious. The survival free of "symptomatic femoral radiological abnormalities" was 93% (95% CI: 89-97) at 5years and 84% (95% CI: 75-95) at 8years. The survival free of "revision for aseptic stem non-integration" was 98% (95% CI: 96.8-100) at 5years and 97% (95% CI: 95.2-100) at 8years. Discussion: In this study, the risk factors for symptomatic radiological abnormalities were being less than 70years of age, having a small lateralized stem and restoring a large femoral offset. Lateralized stems used in this study had a 10% rate of symptomatic radiological abnormalities and a 4% rate of revision for aseptic non-integration. Level of evidence: IV, retrospective study.