Aseptic Loosening in Total Hip Arthroplasty

Abstract

A 51 year old lady with a bilateral hip avascular necrosis of both hips secondary to chronic steroid use for the treatment of systemic lupus erythematosus (SLE) presented with loosening of her right hip implant. The loosening occurred after 16 years of surgery.

Keywords: Aseptic loosening, total hip arthroplasty, systemic lupus erythematosus.

Introduction

Aseptic loosening of orthopaedic implants occurs in the absence of clinical signs of infection. Aseptic loosening of prosthetic components is the major complication of total hip arthroplasty (THA). It is associated with wear particle generation that initiates a proinflammatory cascade, leading to increased osteoclast differentiation and local osteolysis.¹

Case Report (M040027)

A case of 51 year old lady who was diagnosed with systemic lupus erythematous (SLE) in 1983. She was periodically treated with long term steroid therapy. Consequently she developed avascular necrosis (AVN) of both hips. She underwent total hip arthroplasty in 1993 for the right hip AVN with Charnley monobloc prosthesis. In 1997 she underwent second total hip arthroplasty for her left hip.

She was presented to us with complained of pain of right hip since past few months prior to first consultation. She was working as a nurse in a government hospital. The hip pain was affecting her daily living and working activity. The pain mostly occurs during weight bearing or at certain position of the hip. Her medical conditions (SLE, hypertension) is currently under control. She is currently not taking any form of steroid treatments.

Physical examination revealed an antalgic gait. There was a 4 centimeters true limb length shortening. The range of movements of the right hip was full. Other physical examinations were insignificant. The pelvic radiograph of the pelvis showed >2 mm areas of radiolucency in all three De Lee and Charnley zone in the right hip. Protrusio acetabulum was also evidenced in the right hip (Figure 1). Blood parameters did not reveal any sign of infection.

Figure 1: Pelvic radiograph showing areas of osteolysis (black arrows) in the right hip.

A revision surgery was performed on 10/9/2009. The acetabular cup was easily removed. The femoral stem was firmly fixed to the femur. The removal of the stem was fairly straight forward. The cement mantle was carefully removed with an ultrasonic device. After preparation of the acetabular wall, an antiprotrusio mesh was inserted and was secured in placed with two cortical screws. A bone allograft was packed firmly on the acetabular wall. A cemented acetabular cup was inserted followed by a cemented femoral component (Figure 2). A larger offset femoral neck was used to address the leg length discrepancy.

Figure 2: A pelvis radiograph showing right hip revision THA with an anti protrusion mesh in-situ.

Discussion

Total hip arthroplasty (THA) is one of the most successful and effective procedures developed for treatment of pain associated with end-stage arthritis. Long-term follow-up studies report a survivorship greater than 80% at 20 years after surgery.^2,3 As a result of this success, there has been an increasing number of prostheses implanted into younger, more active individuals. This increased functional demand is coupled with increased implant wear rates with standard bearing materials. However, numerous failure mechanisms limit the long-term success including aseptic osteolysis, aseptic loosening, infection, and implant instability.

Aseptic loosening of prosthetic component has been observed since early 1960s by Charnley. He recognized this process when the polytetrafluoroethylene (PFTE) acetabular component he used required revision within 1 to 3 years. A dramatic decreased was seen after switching to polyethylene cup. In those that required revision, he found granulomatous tissues laden with inflammatory cell surrounding the implant at the cement-bone interface. A histopathological analysis by Wilbert and Semlitsch showed that there was excess wear debris within macrophages of periprosthetic tissue. They suggested that the wear debris was biologically active and induced a macrophage response in the tissue surrounding the implant.⁴ Investigators have shown that the cellular activity within this macrophage complex produces a variety of cytokines, enzymes, prostaglandins, and other mediators that stimulate osteoclastic bone resorption and fibrous tissue formation. During the early days of total joint arthroplasty, prosthetic fixation was accomplished through cementing techniques and osteolysis was associated with fragmentation of the cement mantle. Many investigators thus hypothesized that polymethylmethacrylate (PMMA) bone cement wear debris was the cause of osteolysis and aseptic loosening, leading to the term cement disease.⁵ Therefore, improvements in cementing techniques and development of uncemented fixation were emphasized as a means to reduce the incidence of aseptic loosening. However the problem with osteolysis and aseptic loosening persisted, which has led the clinicians and researchers to propose that wear debris from other sources, such as ultra high molecular weight polyethylene (UHMWPE) and metal, as potential causes of osteolysis and subsequent aseptic loosening.

Particulate UHMWPE is the predominant wear debris representing 70-95% of the debris burden found around total hip components.⁶ Wear can occur at the articulating and nonarticulating surface of the UHMWPE acetabular liner. The unfilled screw holes in the metal backing of the acetabular component can become repositories of wear debris. Other source of wear debris includes metal screws that had been used to provide initial cup stability. Fretting corrosion at the modular head and neck junction of the femoral component is another important source of wear debris. Around joint replacements, debris can also be generated from unusual sources such as from degradation of the hydroxyapatite coatings,⁷ fragmentation of cables used for trochanteric fixation,⁸ as well as residuals from the grit blasting and polishing of the components.⁹

Incidence

The incidence of focal osteolysis in cemented cups has been reported from 0 to 19%, with loss of fixation occurring at rates between 0 and 44% at 8 to 10 years. Acetabular components inserted without cement have shown better results at 5 to 7 years than those implanted with cement. Maloney et al have shown that osteolysis developed in 22% of hip replacements in patients younger than 50 years of age and only 7.85% in patients older than 50 years of age.¹⁰

Threaded cups have also produced high rates of osteolysis and osteolysis associated loosening after only 6 years.¹¹ It is now no longer used in the United States due to their high rate of failure.¹² Femoral components inserted with the first-generation cementing techniques loosened at rates between 11 and 30% at 10 years.¹³ However, second and third generation cementing technique have yielded excellent results in terms of osteolysis, loosening, and revision rates. The incidence of osteolysis in femurs with first generation uncemented stems is also quite high. The prosthesis was associated with focal osteolysis rates of 13–52% over short to intermediate follow-up periods.¹³

Pathophysiology

Wear debris varied in sizes and morphology depending on implant design. The shape ranging from spheroids to fibrils and in size from 0.1um to several millimeters.¹⁴ Macrophages appear most sensitive to submicron-sized particles, with giant cells forming around larger (larger than 10 µm) particle.¹⁵

Skeletal bone undergoes continuous remodeling by a process that involves the resorption of bone by osteoclasts coupled with the synthesis and deposition of new bone matrix by osteoblasts. Imbalances between osteoclast and osteoblast activity arise as a consequence of disturbance of controlling cytokines, which can lead to a localized or systemic reduction in bone mass.¹⁶

Aseptic failure of a hip prosthesis occurs as a consequence of a chronic, granulomatous, inflammatory response, which results in the formation of a pseudomembrane at the bone-cement-prosthesis interface.¹⁷ Cytokines, chemokines, and growth factors has been isolated from the pseudomembrane that surrounds failed implants. These include TNF-[alpha], IL-1[beta], IL-6, IL-8, IL-11, TGF-[beta], PGE₂, CSF-1, granulocyte-macrophage colony-stimulating factor (GM-CSF), and matrix metalloproteinases (MMPs). These factors may stimulate osteoclast differentiation and maturation (TNF-[alpha], IL-1) or may be responsible for bone matrix degradation (MMPs).

Although aseptic loosening occurs in the absence of clinical or microbiologic evidence of infection, the possibility of subclinical levels of infection, especially in patients who are immunocompromised is significant.¹⁸ Numerous cell culture and animal model studies support the concept that bacterial endotoxins may contribute to the proinflammatory process and the resultant implant loosening even in the absence of any evidence of infection.^19,20 The potential sources of bacterial endotoxins may come from the manufacturing process and systemic lipopolysaccharide derived from gut flora, minor infections, or dental procedures may adhere to wear particles after they are generated in patients.

Clinical manifestation

Aseptic loosening is defined as an area of radiolucent line around a previously well-fixed prosthesis.²¹ When the zone of radiolucency is wider than 2 mm or becomes circumferential, the prosthesis is considered to be at risk for loosening.²² The radiolucency appears to progress from the intra-articular margin of the interface until it extends circumferentially.

With cemented acetabular components, joint fluid and wear particles tend to deposit at the cement–bone interface. The soft tissue membrane created by the inflammatory response to wear particles proceeds along the cement–bone interface, often leading to its damage. As the interfacial disruption progresses to the acetabular dome, fixation can be lost. A linear osteolytic pattern with the radiolucency occurring at the cement-bone interface can be seen on the radiograph. However, the component tends to migrate into the radiolucent areas in the superior and medial aspects of the acetabulum and may obliterate the radiolucent line. Thus a distinct zone will be seen in the inferior aspects of the socket.

Aseptic loosening in a cemented femoral component has been attributed to joint fluid migration alongside the cement-metal or cement-bone interface. Linear and focal or expansile lesions all have been noted with both the stable and unstable stems.^23,24 Fluid and particles are driven along the interface by the relatively high intra-articular pressures generated during normal gait, which may reach the cement–bone interface via defects in the cement mantle.²⁵ The resulting biological reaction can lead to focal osteolysis in the presence of a well-fixed cemented stem. The 2–3 cm of disruption of the proximal cement–bone interface due to linear osteolysis (i.e., membrane formation) will not significantly compromise the femoral component’s stability, whereas 2–3 cm of disruption at the cement–bone interface in the acetabular socket is likely to have a significant effect on implant stability.

In the cementless acetabular component, the pattern of osteolysis often depends on whether bone ingrowth has occurred. If the socket is stabilized by bone ingrowth, the path of least resistance is via areas that do not have bone ingrowth or via screw holes, which allow particles to migrate into the trabecular bone of the ilium, ischium, and pubis. High particle loads in these areas may be more likely to result in rapidly growing, expansive lesions with indirect margins, leading to osteolysis and progressive bone loss. Loosening often does not result until bone loss is extensive; thus implant failure may be acute and catastrophic, with the patient remaining clinically asymptomatic until the component has loosened. It is important to note that radiographs consistently underestimate the size of osteolytic lesions. The patterns of osteolysis around uncemented cups also depend on the cup design. Regardless of the cup type, both linear and expansile osteolysis has been noted about the fixation screws.^26,27

In the case of cementless stems, the implant design is a major factor in the location of osteolytic lesions. The patch-porous-coated implants are vulnerable to diaphyseal osteolysis as the synovial fluid flows along the smooth portion of the stem into the diaphysis of the bone.²⁸ With circumferential porous-coated stems such as the anatomical medulary locking (AML) or the porous coated anatomic (PCA), focal osteolytic lesions occur most commonly in the proximal aspect of the femur. Occasionally, these lesions have presented as spontaneous fractures of the greater and lesser trochanter.²⁹

Treatment

The treatment and prevention remain challenging to arthroplastic surgeons. The finding of osteolysis must be discussed with patients, and close follow-up is necessary. The factors causing the development of lesions should be explored with each individual patient. Symptoms, location, and the likelihood of progression of the lesions; the amount of bone loss; and the status of fixation of the components must be considered. The surgical procedure must eliminate the osteolytic lesion and the particle generators (the sources of increased wear). Surgery would address removal of granulomatous material when appropriate, filling of the defects with a graft material, and the possible exchange of components, depending on the extent of damage.

Femoral osteolysis

A circumferential osteolysis around cemented stem involving all Gruen femoral zones (figure) can lead to implant loosening if left untreated. Patients need to be follow-up closely at 3 to 6 months interval until the patient is symptomatic or progressive lysis is observed. Operative intervention may be instituted if structural stability is affected. Conditions that may threaten stability include loss of the proximal femoral cortical bone and large osteolytic lesions that may lead to periprosthetic fracture, such as those at the tip of the stem.³⁰

Loose component usually needs to be replaced to prevent further bone resorption. However if the component is well fixed, curettage and grafting of the lytic lesion and retention of the stem, with exchange of the polyethylene liner, may be appropriate.³¹

Cemented stems are considered to be definitely loose by evidence of migration or subsidence of the component, stem debonding, stem fracture or bending, or cement fracture.³² For cementless components, migration and subsidence are most indicative of a loose stem. Additional signs of instability include cortical or cancellous hypertrophy at the stem tip, distal pedestal formation, shedding of the porous coating, and radiolucent lines along the porous coating.³³ Other factors to be considered include the size and location of the lesion as well as patient age, activity, and medical status.

For proximal focal osteolysis, if the component is porous-coated circumferentially, the lesion can be packed with particulate graft and the particle generators exchanged. For osteolysis distal to zones 1 and 7, the femoral stem and polyethylene liner should usually be changed. Cavitary lesions distal to the stem tip present a high risk of periprosthetic fracture and also should be revised with a stem that bypasses the lesion.

Bourne and Rorabeck classified femoral osteolysis into four types and suggested possible treatment modalities in accordance with this classification (Table 1).³⁴

Type	Pathology	Treatment option
1	Intact cancellous bone and cortical tube and well-preserved cancellous bone.	Can use any primary stem (cemented or uncemented)
2	Deficient cancellous bone and and an intact cortical tube	Porous-coated prostheses. Proximal porous coated prostheses. Impaction allografting with cemented prostheses (<65 years of age )
3	Deficient cancellous bone and cortical tube	Similar to type 2. Distal fixation with a scratch fit of 4–6 cm is more difficult to obtain and cortical strut allograft is needed in many cases.
4	Absent cancellous bone and cortical tube	Proximal femoral allograf. Tumor prosthesis.

Table 1: Classification of femoral osteolysis and treatment option.

Figure 3: Treatment algorithm for femoral osteolysis. (Rubashetal. Osteolysis:Surgical treatment. In:Cannon WD, ed. Instructional Course Lecture 47. Rosement, IL:American Academy of Orthopaedic Surgeons, 1997:321–330.)

Acetabular osteolysis

A systematic approach to the treatment of acetabular osteolysis may follow the algorithm formulated by Rubash et al.³⁵ An evaluation of implant stability is important for cemented cups. Linear or focal osteolysis in two or three acetabular zones has been associated with 71 and 94% incidences of loosening of the component, respectively.³⁶ A loose-cemented component must be revised, preferably with porous-coated shell. Bone stock deficiencies must be treated with appropriate grafting. If the cup is not loose, the degree of wear should be evaluated. Worn cups with eccentricity of the femoral head should be replaced.

Uncemented cups have been classified into three separate groups. The type I cup is stable with discrete focal osteolysis, including zones 1 and 3, and it is occasionally adjacent to screws. The component usually can be retained, and particulate graft can be packed into the defect if readily accessible. Type II components are also stable by virtue of bone ingrowth, but the function of the cup is compromised by damaged locking mechanism of a modular cup, extensive wear of the shell, or malpositioned shell. In these cases, the entire component is often removed, defects are filled with the appropriate graft, and a new cup is reimplanted without cement. Type III cups are unstable and have migrated into the osteolytic lesion, necessitating exchange.

Figure 4: Treatment algorithm for acetabular osteolysis. (Rubashetal. AAOS 1997)

Non surgical Treatment

As alternatives to surgical intervention, investigators have looked to modifying the host response to debris by using nonsteroidal anti-inflammatory drugs to inactivate the inflammatory mediators. Modulation of the actions of anti-TNF-(alpha) has been explored as a potential treatment for particle-induced osteolysis. Etanercept, a soluble inhibitor of TNF-[alpha], inhibits osteoclastic bone resorption in a bone wafer pit assay and cytokine production from titanium particle-stimulated macrophages.³⁷ The bisphosphonates are a class of drug that are analogs of pyrophosphate. They are commonly used for medical treatment of osteoporosis, Paget's disease, and hypercalcemia. Animal model studies suggest these agents may be of potential benefit for treatment of particle-induced osteolysis.^38,39 Inhibition of RANK-RANKL interactions has been effective in blocking bone loss in animal models of particle-induced osteolysis. AMG-162 is a human immunoglobulin G₂ monoclonal antibody with a very high affinity for human RANKL. The safety and bone antiresorptive effects of AMG-162 in postmenopausal women has been observed in Phase I and II clinical trials.^40,41 This drug may offer a future therapeutic or preventative agent for particle-induced osteolysis.

Conclusion

Aseptic loosening is a major complication in total hip arthroplasty. The hip articulation, regardless of the bearing coupling, is particle generators which will release wear debris into the joint and the surrounding area which then will activate cellular defense mechanism. Improved models of osteolysis will enable more detailed study of particular treatment methods such as improved grafting materials and techniques, pharmacologic intervention, or immunologic modulation. With these advances, it is now possible to expect joint replacements to last more than 20 years.

Reference

1. Greenfield EM, Bi Y, Ragab AA, Goldberg VM, Van De Motter RR. The role of osteoclast differentiation in aseptic loosening. J Orthop Res. 2002;20:1-8.
2. Berry DJ, Harmsen WS, Cabanela ME, Morrey BF. Twenty-five-year survivorship of two thousand consecutive primary Charnley total hip replacements. J Bone Joint Surg Am. 2002;84:171-177.
3. Callaghan JJ, Albright JC, Goetz DD, Olejniczak JP, Johnston RC. Charnley total hip arthroplasty with cement: minimum twenty-five-year follow-up. J Bone Joint Surg Am. 2000;82:487-497
4. Willert HG, Semlitsch M. Reactions of the articular capsule to wear products of artificial joint prostheses. J Biomed Mater Res. 1977;11:157–164.
5. Jones LC, Hungerford DS. Cement disease. Clin Orthop. 1987;225:192–206.
6. Maloney WJ, Smith RL, Schmalzried TP, Chiba J, Huene D, Rubash HE. Isolation and characterization of wear particles generated in patients who have had failure of a hip arthroplasty without cement. J Bone Joint Surg. 1998;77-A: 1301–1310
7. Morscher EW, Hefti A, Aebi U. Severe osteolysis after third-body wear due to hydroxyapatite particles from acetabular cup coating. J Bone Joint Surg. 1998;80-B:267–272.
8. Hop JD, Callaghan JJ, Olejniczak JP, Pedersen DR, Brown TD, Johnston RC. The Frank Stinchfield Award. Contribution of cable debris generation to accelerated polyethylene wear. Clin Orthop. 1997;344:20–32.
9. Merchant KK, Rohr WL, Lintner WP, Bhambri SK. Orthopaedic implant surface debris. Trans Orthop Res Soc. 1995;20:164–164.
10. Maloney WJ, Galante JO, Anderson M, et al. Fixation, polyethylene wear, and pelvic osteolysis in primary total hip replacement. Clin Orthop. 1999;369:157–164.
11. Fox GM, McBeath AA, Heiner JP. Hip replacement with a threaded acetabular cup. A follow-up study. J Bone Joint Surg. 1994;76-A:195–201.
12. Bruijn JD, Seelen JL, Feenstra RM, Hansen BE, Bernoski FP. Failure of the Mecring screw-ring acetabular component in total hip arthroplasty. A three to seven-year follow-up study. J Bone Joint Surg. 1995;77-A:760–766.
13. Sinha RK, Shanbhag AS, Maloney WJ, Hasselman CT, Rubash HE. Osteolysis: cause and effect. Instr Course Lect. 1998;47:307–320.
14. Campbell P, Ma S, Yeom B, McKellop H, Schmalzried TP, Amstutz HC. Isolation of predominantly submicron-sized UHMWPE wear particles from periprosthetic tissues. J Biomed Mater Res. 1995;29:127-131
15. Maloney WJ, Smith RL, Hvene D, Schmalzried TP, Rubash H. Isolation and characterization of wear debris generated in patients who have had failure of a hip arthroplasty without cement. J Bone Joint Surg Am. 1994;77:1301-131
16. Mostov K, Werb Z. Journey across the osteoclast. Science. 1997;276:219-220.
17. Goldring SR, Schiller AL, Roelke M, Rourke CM, O'Neil DA, Harris WH. The synovial-like membrane at the bone-cement interface in loose total hip replacements and its proposed role in bone lysis. J Bone Joint Surg Am. 1983;65:575-584
18. Greenfield EM, Bi Y,Ragab AA. Goldberg VM, Nalepka JL, Seabold JM. Does endotoxin contribute to aseptic loosening of orthopedic implants? J Biomed Mater Res B Appl Biomater. 2005;72: 179-185
19. Bi Y, Collier TO, Goldberg VM, Anderson JM, Greenfield EM. Adherent endotoxin mediates biological responses of titanium particles without stimulating their phagocytosis. J Orthop Res. 2002;20:696-703
20. Daniels AU, Barnes FH, Charlebois SJ, Smith RA. Macrophage cytokine response to particles and lipopolysaccharide in vitro. J Biomed Mater Res. 2000;49:469-478
21. Stauffer RN. Ten-year follow-up study of total hip replacement. J Bone Joint Surg. 1982;64-A:983–990.
22. Kim YH, Kim VE. Uncemented porous-coated anatomic total hip replacement. Results at six years in a consecutive series. J Bone Joint Surg. 1993;75-B:613.
23. Maloney WJ, Jasty MJ, Harris WH, Galante JO, Callaghan JJ. Endosteal erosion in association with stable uncemented femoral components. J Bone Joint Surg. 72-A:1025–1034, 1990.
24. Zicat B, Engh CA, Gokcen E. Patterns of osteolysis around total hip components inserted with and without cement. J Bone Joint Surg .1995;77-A:432–439.
25. Anthony PP, Gie GA, Howie CR, Ling RSM. Localised endosteal bone lysis in relation to the femoral components of cemented total hip arthroplasties. J Bone Joint Surg. 1990;72-B:971–979.
26. Santavirta S, Konttinen YT, Bergroth V, Eskola A, Tallroth K, Lindholm TS. Aggressive granulomatous lesions associated with hip arthroplasty. J Bone Joint Surg. 1990;72-A:252–258.
27. Fox GM, McBeath AA, Heiner JP. Hip replacement with a threaded acetabular cup. A follow-up study. J Bone Joint Surg. 1994;76-A:195–201.
28. Bobyn JD, Jacobs JJ, Tanzer M, Urban RM, Aribindi R, Sumner DR, Turner TM, Brooks CE. The susceptibility of smooth implant surfaces to periimplant fibrosis and migration of polyethylene wear debris. Clin Orthop. 1995;311:21–39.
29. Heekin RD, Engh CA, Herzwurm PJ. Fractures through cystic lesions of the greater trochanter. A cause of late pain after cementless total hip arthroplasty. J Arthrop. 1996;11:757–760.
30. Rubash HE, Sinha RK, Shanbhag AS, Kim SY. Pathogenesis of bone loss after total hip arthroplasty. Orthop Clin North Am. 29:173–186, 1998
31. Schmalzried TP. Management of pelvic osteolysis. Transactions of the 26^th Annual Hip Course. Boston: Harvard Medical School 1996.
32. Harris WH, Schiller AL, Scholler JM, Freiberg RA, Scott R. Extensive localized bone resorption in the femur following total hip replacement. J Bone Joint Surg. 1976;58-A:612–618.
33. Callaghan JJ, Salvati EA, Pellicci PM, Wilson PD, Jr., Ranawat CS. Results of revision for mechanical failure after cemented total hip replacement, 1979 to 1982. A two- to five-year follow-up. J Bone Joint Surg. 1985;67-A:1074–1085.
34. Dunbar MJ, Blackley HR, Bourne RB. Osteolysis of the femur: principles of management. Instr Course Lect. 2001;50:197–209.
35. Sinha RK, Maloney W.J, Paprosky WG, Rubash HE. Surgical treatment of osteolysis. In: Callaghan JJ, Rosenberg AG, Rubash HE, eds. Adult Hip. Philadelphia-New York: Lippincott-Raven. 2001;1998:1549–1570.
36. Hodgkinson JP, Shelley P, Wroblewski BM. The correlation between the roentgenographic appearance and operative findings at the bone-cement junction of the socket in Charnley low friction arthroplasties. Clin Orthop. 1988;228: 105–109.
37. Childs LM, Goater JJ, O'Keefe RJ, Schwarz EM. Effect of antitumor necrosis factor-alpha gene therapy on wear debris-induced osteolysis. J Bone Joint Surg Am. 2001;83:1789-1797
38. Horowitz SM, Algan SA, Purdon MA. Pharmacologic inhibition of particulate-induced bone resorption. J Biomed Mater Res. 1996;31:91-96.
39. Thadani PJ, Waxman B, Sladek E, Barmada R, Gonzalez MH. Inhibition of particulate debris-induced osteolysis by alendronate in a rat model. Orthopedic. 2002;25:59-63.
40. Bekker PJ, Holloway DL, Rasmussen AS, et al. A single-dose placebo-controlled study of AMG 162, a fully human monoclonal antibody to RANKL, in postmenopausal women. J Bone Miner Res. 2005;20:2275-2282.
41. McClung MR, Lewiecki EM, Cohen SB, Bolognese MA, et al. AMG 162 Bone Loss Study Group. Denosumab in postmenopausal women with low bone mineral density. N Engl J Med. 2006;354:821-831.