© The Author(s) 2021
Article reuse guidelines:
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
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,
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-
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 width (mm)
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
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).
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
4 HIP International 00(0)
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)).
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)).
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
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–
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.
Analysis of classification agreement showed very good
inter-observer and good intra-observer agreement for both
undersizing and the sub-type categorisation of undersizing
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.
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
Champagne Fluted Undersized
Figure 6. Classification of undersizing and mix of undersizing subtypes.
Table 3. Observer agreement for undersizing classification
Agreement of sub-
type of undersizing
Inter-observer Kappa Very good (0.826) Very good (0.818)
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)
(cancellous space greater
than tip width)
Table 5. Digital templating compared to varus undersized
undersized < 2
undersized ⩾ 2
Not Undersized (triangle
overlap < 50%)
Varus Undersized (triangle
overlap > 50%)
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.
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.
% 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;
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;
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.