Knee Surgery, Sports Traumatology, Arthroscopy

, Volume 20, Issue 3, pp 487–494

Improved stability with intramedullary stem after anterior femoral notching in total knee arthroplasty

Authors

    • Department of Mechanical EngineeringUniversity of Aveiro
  • F. Fonseca
    • Department of OrthopedicsCoimbra University Hospital
  • C. Relvas
    • Department of Mechanical EngineeringUniversity of Aveiro
  • A. Ramos
    • Department of Mechanical EngineeringUniversity of Aveiro
  • J. A. Simões
    • Department of Mechanical EngineeringUniversity of Aveiro
Knee

DOI: 10.1007/s00167-011-1557-2

Cite this article as:
Completo, A., Fonseca, F., Relvas, C. et al. Knee Surg Sports Traumatol Arthrosc (2012) 20: 487. doi:10.1007/s00167-011-1557-2

Abstract

Purpose

It has been hypothesized that femoral notching in total knee arthroplasties weakens the cortex of the femur, which can predispose to femoral fractures in the postoperative period. Some authors suggest that patients who sustain inadvertent notching should have additional protection in the postoperative period, and consideration should be given to the use of prophylactic femoral stems. In this case, a question can be raised: Is the use of femoral stem in an anterior femoral notching an effective way to reduce the fracture risk? We hypothesized that for a larger notch, the use of a femoral stem does not decrease considerably the stress-riser at the notch edge, and the use of stem is not enough to reduce the risk of fracture.

Methods

In the present in vitro study, twelve synthetic femurs were selected and used for the experiments under two load scenarios. Femoral components with and without femoral stems were implanted in femurs with different notch sizes to predict experimentally the strain levels at the notch edge with the use of fiber Bragg gratings and at notch region with strain gauges.

Results

Despite the global strain reduction in stemmed condition, at the notch edge, the strain behavior was dissimilar for the different notch depths. For notch depths lower than 5 mm, the use of stem reduces the strain level at the notch edge to values below the intact femur condition, while for depths greater or equal to 5 mm, the strain levels at the notch edge were higher than the intact femur condition with values ranging from +10 to +189%.

Conclusions

The present study suggests the use of a prophylactic stem for notch depths greater than 5 mm. For notch depths below 5 mm, the fracture risk due to strain increase at the notch edge seems to be low in the stemless condition.

Keywords

Femoral notchingPeriprosthetic femur fractureExperimental studyIntramedullary stemTotal knee arthroplasty

Introduction

The occurrence of distal femoral fracture after total knee arthroplasties (TKA) is approximately 1% with reported rates ranging from 0.3 to 2.5% [15]. Documented risk factors are osteopenia, osteoporosis, rheumatoid arthritis, neurologic disorders, chronic steroid therapy, and anterior femoral notching [112]. Among the several factors associated with these fractures, osteopenia or conditions that induce osteopenia and femoral notching are implicated most often [1, 3, 6, 1316]. The prevalence of notching of the anterior femoral cortex in TKA has been reported to range from 3.5 to 41% [1, 3, 4]. The possible role of notching of the anterior femoral cortex in distal femoral fractures following TKR has been observed clinically and studied biomechanically [1, 6, 9, 15, 1719]. Hirsh et al. [6] hypothesized that femoral notching weakens the cortex of the femur, which can predispose to femoral fractures in the early postoperative period. The incidence of supracondylar fracture in a notched femur following TKR varies from 0.5 to 44.3% [16, 12, 15]. Although some authors refer no relationship between minimal anterior femoral notching and supracondylar fracture of femur [1, 4, 12], others conclude biomechanically that notching reduces the femur strength [17, 18]. Lesh et al. [17] investigated the biomechanical effects of notching of the distal anterior femoral cortex in TKR using human cadaveric femurs and found a mean reduction of 18% in bending strength and a 42% mean decrease in torsional strength for full-thickness notching of the femur. Shawen et al. [18] also used cadaveric femora to demonstrate that a 3-mm anterior cortical notch reduces torsional load to failure. Lesh et al. [17] advocates that patients who sustain inadvertent notching should have additional protection at postoperative period, and consideration should be given to the use of a femoral stem extension as a mean to bypass the stress-riser of the anterior notch. However, the use of prophylactic femoral stems intraoperatively is always more complex and expensive. In this case, a question can be raised: For which notch size does the use of a femoral stem reduce the fracture risk? Thus, the aim of the present study was to determine the range of notch size for which the use of a prophylactic femoral stem becomes advantageous to reduce the risk of fracture. For the present study, implanted femurs were used to predict experimentally the strain levels at the notch edge with fiber Bragg gratings and at the notch region with strain gauges. We hypothesized that for small notch sizes, the use of a prophylactic stem is unnecessary, and for a larger notch, the use of a femoral stems does not decrease considerably the stress-riser at the notch edge and the use of stem is not enough to reduce the fracture of risk.

Materials and methods

Twelve synthetic femurs (left, mod. 3306 from Pacific Research Labs, Vashon Island, WA, USA) were selected and used for the experiments. The physical construction of the this type of synthetic bone has recently been shown to reveal no statistically significant difference in both structural properties and strains when compared with the same obtained with cadaver bones, exhibiting extremely low specimen-to-specimen variability [20, 21]. Femoral components with and without femoral stems of the P.F.C Sygma Modular Knee System (DePuy International, Inc) were implanted in synthetic femurs (Fig. 1). Six femurs were implanted with a femoral component, and the other six femurs were implanted with a femoral component and a press-fit stem extension (Table 1). The relative positions between femur and femoral component were identical in all femurs of the two groups, which were achieved by the use of a set of jigs used for the bone cuts. Each of these groups was separated in two sub-groups of 3 femurs. In one sub-group, the notch defects were produced at 2 mm from the prosthesis and in the other sub-group at 4 mm (Fig. 1). In all femurs, seven different notch depths ranged 1–7 mm were produced. These notch depths were successively increased (1 mm) from the first experiment without a notch to the final depth of 7 mm. In all femurs, the radius notch was 0.7 mm, similar to what is obtained with a standard saw in knee surgery. The distances between the notch and femoral component as well the notch depths were considered after the observation of 296 knee radiographs (215 women and 81 men, with a mean age of 67 years, Orthopaedic Department, Coimbra Medical School, Coimbra University Hospital), including 36 knees (12%) with femoral notching (Fig. 2). The maximum notch depth found was 7.7 mm with a mean depth of 3.5 mm.
https://static-content.springer.com/image/art%3A10.1007%2Fs00167-011-1557-2/MediaObjects/167_2011_1557_Fig1_HTML.gif
Fig. 1

Pictures of the composite femur models with notch, strain gauges positions (A anterior, M medial, L lateral) and fiber Bragg grating

Table 1

Dimensions and materials of the femoral component and stem used in the implanted femoral constructions

Implant

Dimension P.F.C. knee system

Material

https://static-content.springer.com/image/art%3A10.1007%2Fs00167-011-1557-2/MediaObjects/167_2011_1557_Figa_HTML.jpg

Stabilized femoral component—Size 4 71 mm ML, 65 mm AP sacrifice of LCP

Co-Cr-Mo

https://static-content.springer.com/image/art%3A10.1007%2Fs00167-011-1557-2/MediaObjects/167_2011_1557_Figb_HTML.jpg

Press-fit stem—18 mm × 175 mm

Ti-6AL-4V

https://static-content.springer.com/image/art%3A10.1007%2Fs00167-011-1557-2/MediaObjects/167_2011_1557_Fig2_HTML.jpg
Fig. 2

Radiograph of knee after TKA with anterior femoral notching (left) and schematic representation of notch depth (right)

All femurs were subjected at two load scenarios, which provoke simultaneously bending and torsion loads in the distal femur. For both load scenarios, the femurs were rigidly fixed at the diaphysis at a distance of 300 mm from the condylar surface and at 45° of flexion in the sagittal plane (Fig. 3). In load scenario 1, a load of 1,500 N (controlled by a load cell TC4 1T, AEP, Modena, Italy) was applied at the proximal femur and the femoral component was simply supported at the medial condyle. This load configuration generated a severe bending and torsion effect in the femur. In load scenario 2, a load of 2,990 N was applied in femur, with the femoral component supported by both femoral condyles (Fig. 3). The position of the sphere shifted medially relatively to the middle distance between the condyles, causes a load repartition of 60% in medial condyle and 40% in the lateral condyle. The load magnitudes applied to the synthetic femurs were based on the work of Heinlein et al. [22] and Matthews et al. [23].
https://static-content.springer.com/image/art%3A10.1007%2Fs00167-011-1557-2/MediaObjects/167_2011_1557_Fig3_HTML.gif
Fig. 3

Experimental loading setup used for load case 1 (Left) and load case 2 (Right)

Three triaxial strain gauges (KFG-3-120-D17-11L3M2S, Kyowa Electronic Instruments, Japan) were glued in each femur in the medial (M), lateral (L), and anterior (A) distal cortex, adjacent to the notch region (Fig. 1). The same position for each strain gauge was maintained in the twelve femurs, which were marked using a 3D coordinate measuring machine (Mod. Maxim, Aberlink, U.K.). All strain gauges were connected to a data acquisition system PXI-1050 (National Instruments, USA). In addition, fiber Bragg grating sensors (2.5 mm of length) were used to measure the strains along the cortex cut at the notch edge (0.7 mm of radius) and at 2 mm above the notch (Fig. 1). All fiber Bragg sensors were connected to a data acquisition system BraggScop 1500 (FiberSensing, Portugal), and the measured wave length was converted automatically in a strain value. The measurement accuracy is 1μstrain. In each load experiment, the femur was first pre-conditioned with a pre-load of 500 N. The maximal (ε1) (positive) and minimal (ε2) (negative) principal strains of the strain gauges and the strain of the fiber Bragg gratings were calculated and averaged for each notch depth.

Statistical analysis

For each load scenario and notch depth, the mean of strains was compared with intact situation (unnotched) for the stemmed and stemless condition. Student’s t tests for paired samples were performed, and P values were calculated using a statistical analyses package (SPSS, USA). Statistically significant differences will be detected whether P value <0.05.

Results

The means and standard deviations of strains in the stemless femurs for load scenario 1 and load scenario 2 for the notch at 2 mm of the prosthesis can be depicted from Fig. 4 for notch depths of 1, 3, 5, and 7 mm. The difference of principal strains at the strain gauges positions A (anterior), M (medial), and L (lateral) were inferior to 12% between notches at 2 mm and 4 mm from the prosthesis (proximal edge) for both load scenarios. Moreover, for both load scenarios and strain gauges, the nominal values of the minimal (ε2) (negative) principal strains were greater (1.5–3 times) than the nominal values of the maximal (ε1) (positive) principal strains. The highest minimal principal strain values (ε2) were obtained in the medial (M) strain gauge position for both load scenarios and notch depths. In Table 2, we present the strain differences (nominal value and percentage) between the intact (unnotched stemless) and the notched stemless femurs for load scenario 2 with notches at 2 mm of the prosthesis. The raise of the notch depth until 3–4 mm in the stemless models increased the principal strains at the medial (M) and lateral (L) strain gauges in 3% and a small reduction of less than 2% at the anterior position (A) relatively to unnotched case. For the notch depths between 5 and 7 mm, the increase of the minimal (ε2) principal strains ranged between +20 and +69% in the (M) medial and lateral (L) strain gauges positions; the opposite happened at the anterior (A) position with a decrease that ranged from −32 to −91% for both load scenarios. At the notch edge (fiber Bragg grating), the stemless femurs with the notch at 2 mm from the prosthesis presented mean strain values +24% higher than the femurs with the notch at 4 mm from the prosthesis. The increase of the notch depth until 3–4 mm increased the mean strain at the notch edge (Bragg grating) in +12% relatively the unnotched (stemless) situation. For notch depths between 5 and 7 mm, the strain along the Bragg grating was compressive (negative strain) with an increase relatively to the unnotched situation that ranged from +77% (+400μstrain, load scenario 2) to +411% (+716μstrain, load scenario 1).
https://static-content.springer.com/image/art%3A10.1007%2Fs00167-011-1557-2/MediaObjects/167_2011_1557_Fig4_HTML.gif
Fig. 4

Means and standard deviations of measured principal strains (ε1—maximal and ε2—minimal) at each strain gauge position (A anterior, M medial, L lateral) and strain at the notch edge (fiber Bragg grating) for stemmed and unstemmed femurs for load case 1 (Right) and load case 2 (Left)

Table 2

Nominal and percentage changes of mean strains between intact (unnotched stemless) femur and notched stemless and stemmed femurs at the strain gauges position M, L, and A and at the fiber Bragg grating at the notch edge

Sensor

A

L

M

Bragg grating (μstrain)

Principal strain

ε1 (μstrain)

ε2 (μstrain)

ε1 (μstrain)

ε2 (μstrain)

ε1 (μstrain)

ε2 (μstrain)

Model

Value change

%

Value change

%

Value change

%

Value change

%

Value change

%

Value change

%

Value change

%

Load case 2 (Strain differences between Intact and Notched femurs in stemless condition)

Notch depth 1 mm

−33

−6

−65

−4

62

10

−17

−1

23

7

23

2

53

10

Notch depth 3 mm

11

2

−51

−3

31

5

26

2

46

13

70

5

65

13

Notch depth 5 mm

50

9

−561

−37

−100

−16

251

20

31

8

243

18

400

77

Notch depth 7 mm

111

19

−896

−91

−149

−23

450

36

45

12

449

34

595

114

Load case 2 (Strain differences between Intact and Notched femurs in stemmed condition)

Notch depth 1 mm

−222

−42

−505

−33

−289

−45

−579

−46

−111

−33

−385

−29

−107

−21

Notch depth 3 mm

−196

−37

−447

−29

−291

−46

−540

−43

−65

−19

−365

−28

−44

−8

Notch depth 5 mm

−60

−11

−967

−62

−330

−52

−450

−36

−173

−51

−248

−19

50

10

Notch depth 7 mm

−47

−9

−1,070

−69

−338

−53

−360

−29

−165

−48

−299

−23

83

16

The strain results of stemmed femurs can be depicted in the Fig. 4 for a notch at 2 mm of the prosthesis. As expected, the use of the stemmed femur reduces the strain levels at the strain gauge positions (M, L, and A) and at the notch edge (Bragg grating) for both load scenarios, when compared with the stemless condition. In Table 2, we present the nominal strain differences and percentages between the intact (unnotched stemless) and stemmed femurs for load scenario 2 with the notch at 2 mm of the prosthesis. The highest minimal principal strain (ε2) reduction relatively to the intact femur occurred at the anterior (A) strain gauge with a reduction of −570μstrain (−58%) for load scenario 1 and −1,070μstrain (−69%) for load scenario 2. At the notch edge, the strain reduction ranged from −8 to −21% for the notch depths until 3–4 mm for both load scenarios relatively to the intact femur. For notch depths greater than 5 mm, the use of a stem does not reduced the strain at the notch edge to levels below the one observed for the intact (unnotched stemless) condition, the strain excess ranged from +10% (+50μstrain) to +16% (+83μstrain) for load scenario 2 and from +155% (+269μstrain) to +189% (+328μstrain) for load scenario 1.

The significance level (paired t test for data with equal variance, P < 0.05) of the differences of mean strain at the notch edge (Bragg grating) between notched femurs (different depths) and the intact condition (unnotched, stemless) is presented in Table 2. For both load scenarios and stem condition (stemless or stemmed), no significant differences (P < 0.05) were found for notch depths until 5 mm. Contrary, for notch depths greater than 5 mm, significant differences (P < 0.05) were found for both load scenarios in stemless condition and for load scenario 1 in the stemmed condition.

Discussion

The most important finding of the present study was that the use of a prophylactic stem reduces the strain-riser at the notch edge in anterior femur, contributing to reducing the fracture risk, in particular for larger notch depths, contradicting our study hypothesis. The experimental strain results for both load scenarios were highly repeatable at the different strain gauge positions as well as the strains at the notch edge measured through the fiber Bragg grating. The standard deviations obtained for each construction was less than 12% of the averaged strain value. These values were in range of previous experimental studies where composite bones were used [20, 21, 24]. The main reason to study the changes of strains is supported by the evidence that the onset of fracture in cortical bone is consistent with the strain-based criterion, which has been used in theoretical models of the mechanical bone behavior [25, 26].

The measured strains in the stemless femurs for notch depths from 1 to 5 mm evidenced no significant differences relatively to the unnotched situation, while significant (P < 0.05) strain differences were recorded for notch depths ranged from 5 to 7 mm. Moreover, the strain changes at the notch edge (Bragg sensor) were more pronounced than the changes at the strain gauges with the increase of the notch depth. The strain results also confirmed that the distance between the notch and the proximal femoral component modify the cortex strain patterns. The notch nearest to the femoral component (at 2 mm) presented the highest strains at the notch edge (Bragg grating), while the principal strains (M, L, and A) were only slightly different from the notch at 4 mm from the prosthesis. Opposed structural behaviors were observed at the cortex with the increase of notch depth. Notch depths greater than 5 mm (full-thickness cortical defect) led to a significant (P < 0.05) strain increases at the notch edge (Bragg sensor) and at the medial (M) and lateral (L) strain gauges and a significant (P < 0.05) decrease of principal strains at the anterior (A) strain gauge relatively to the unnotched situation. These strain increases at the notch edge and at the medial and lateral notch sides can be related with the reduction of the resistant area at the distal cortex, due to the notch cut, while the strain reduction observed at the anterior region can be related with the discontinuity (full-thickness notch) of the load transmission between the femoral component and the anterior cortex region. The strain-shielding effect in the anterior femur region (A) can be a factor to promote later bone osteopenia. These two structural cortex behaviors (strain rise and strain-shielding) for notch depths beyond 5 mm are mechanical factors that can potentially contribute to weaken the femur cortex, which can predispose to femoral fractures in the postoperative period [1, 3, 6, 18, 27].

Lesh et al. [17] and Shawen et al. [18] using cadaveric femurs found a decrease of the load to failure in femurs with a notch depth of approximately 3 mm (full-thickness notch), ranged from 39% [18] to 18% [17]. Zalzal et al. [28] in a purely finite element study suggest an increase of the risk of supracondylar fracture for notch depths greater than 3 mm. Moreover, Lesh et al. [17] reported that fractures were originated at the notch edge. The load decrease to failure and the initial region of fracture in full-thickness notched femurs observed in the precedent cadaveric studies [17, 18] do not contradict our experimental results where for notch depths greater than 3 mm, strain rises at the notch edge (Bragg grating) from 13% to 411%, relatively to the unnotched situation, were observed. However, just for notch depths between 5 and 7 mm, we found nominal strain increases of several times of the unnotched situation, this certainly increases the fracture risk in the immediate postoperative period for full weight patient. For notch depths ranging from 3 to 4 mm, the nominal strain increase was small relatively to unnotched situation, and at this case, the fracture risk is in the immediate postoperative period resemble reduced, but this slight strain raise effect at the notch edge combined with an osteopenia bone environment may also contribute to a bone fatigue damage process and consequently can lead to a later bone fracture (stress fracture) [29]. Markel and Johnson [30] reported that seven of eight patients with anterior femoral cortical notching had fracture within 3 months after TKA. The risk associated with notch great depths identified at this experimental study as well in other previous studies [17, 18, 28] can be the reason for the surgeon to consider additional protection in the postoperative period. This protection in the postoperative period can be done by avoiding full weight bearing or through the use of prophylactic femoral stem as a mean to bypass the strain-riser at the notch region. The use of a stem as a prophylactic mean was suggested by Lesh et al. [17]. However, can the femoral stem effectively reduce the fracture risk and their use be considered by the surgeon?

The measured strains for all notch depths in the stemmed femurs were lower than the notched stemless condition. At the strain gauges (A, M, and L), a decrease of principal strains ranged from −9% to −53% occurred for all notch depths relatively to the intact situation. Despite the significant (P < 0.05) strain reduction relatively to intact condition in all strain gauges, at the notch edge (Bragg grating), the strain behavior was dissimilar for the different notch depths. For notches depths lower than 5 mm, the use of stem not reduces significantly (P < 0.05) the strain at the notch edge to values below the intact femur condition, while for depths greater or equal to 5 mm, the strain levels at the notch edge (Bragg sensor) were significantly higher (P < 0.05) than the intact femur condition with values ranging from +10% (50μstrain) in load scenario 2 to +189% (328μstrain) in load scenario 1; nevertheless, the strain rise at the notch edge was reduced until 50% relatively to the stemless condition. This contradicts our initial hypothesis that for largest notch depths, the use of femoral stem does not decrease the fracture risk. Thus, the uses of a stem in notched femurs have contradicted effects. In fact, the stem use reduces globally the strain at the notch region, and this effect is very positive at the notch edge, reducing the risk of fracture when compared with the notched stemless condition, but on the other hand, the generalized strain-shielding effect in the diaphysis region relatively to the intact condition (unnotched stemless) can contribute for the osteopenia mechanism, which is a risk factor associated with femur fracture after TKA. Therefore, the use prophylactic stem in notched femur should take into account the advantage and disadvantages described previously, as well as the level of osteopenia at the femur (Table 3).
Table 3

P values obtained from t tests, performed to test the difference of mean of strains in notch edge (Bragg sensor) between notched and the intact femur (unnotched) at the stemless and stemmed conditions, for both load cases

Notch depth

P values

Load case 1

Load case 2

Stemless

Stemmed

Stemless

Stemmed

1 mm

0.26

0.18

0.19

0.16

3 mm

0.24

0.21

0.22

0.27

5 mm

0.03 (P < 0.05)

0.04 (P < 0.05)

0.03 (P < 0.05)

0.28

7 mm

0.02 (P < 0.05)

0.02 (P < 0.05)

0.01 (P < 0.05)

0.31

As in all experimental studies, the present study had some shortcomings. One limitation of this study is concerned with the use of synthetic bones and simplifications of the experiments to represent the functioning knee. The advantage of these analogue models is that their variability is significantly lower than that of cadaveric specimens for all loading regimens [20, 21]. In addition, the flexural and torsional rigidity of synthetic femur are within range of healthy adult bones (age <80 years old), as well as the failure modes of this synthetic models were close to published findings for human bones [21]. Other limitation was the experimental load cases that were simplified in terms of the applied loads and the structural links (ligamentous, muscles, etc.) in knee; however, the applied load cases are representative of the major loads acting upon the implant and bone structure. Another restriction of this study is related with the bone quality, the nominal results observed are representatives of healthy adult bones [21], and the nominal results might be different with more pathological bone. It should be also referred that other studies including different femoral component and stem geometries may lead to slightly different results. However, due to the comparative nature of the study, these limitations would not change the major structural differences between stemless and stemmed models in the presence of different notch depths.

Many times inside operation, surgeon is faced with a notching that he does not want but happens. At that moment, he needs an answer to the questions, did I need to put a stem to prevent fracture or I didn’t. Should I continue the surgery without changing plans? Is sure that my patient will not have a fracture? The intent of this study was give to the clinicians, insights into the helpfulness or not of the use of prophylactic femoral stems in the reduction of the risk of femoral fracture due to the inadvertent notching in the anterior femur, and therefore predict suitable surgical plans. The study suggests a higher fracture risk at the postoperative period based in the significant strain rise at notch edge for notch depths greater or equal to 5 mm. Therefore, maybe is prudent clinically the use of a prophylactic stem in this notch condition, where the strain rise at the notch edge is reduced until 50% relatively to the stemless condition. For notch depths below 5 mm, the fracture risk due to strain rise at the notch edge seems to be low. In this case, it may be prudent to avoid the use of a stem and avoid full weight bearing during the postoperative period.

Conclusions

The present study suggests the use of a prophylactic stem for notch depths greater than 5 mm, due to the fracture risk, related with the significant strain rise at the notch edge. For notches depths below 5 mm, the strain rise at the notch edge seems to be low and the stem use not presents major structural advantages, being to avoid.

Copyright information

© Springer-Verlag 2011