Suboptimal (thin) distal cement mantle thickness as a contributory factor in total hip arthroplasty femoral component failure

The Journal

of Arthroplasty

Vol. 9 No. 2 1994

Suboptimal (Thin) Distal Cement Mantle Thickness as a Contributory Factor in Total Hip Arthroplasty Femoral Component Failure A Retrospective Radiographic Analysis Favoring Distal Stem Centralization Michael J. Star, MD, Clifford W. Colwell, Jr., MD, Gary J. Kelman, MD, R. Tracy Ballock, MD, and Richard H. Walker, MD

One hundred cemented total hip arthrcsplasties(THAs) were evaluated regardingthe potential benefit of THA femoral component distal stemcentralization, specificallyregardingcementmantle thickness.Factorspotentially predictive of f’emoral component mechanicalIr)osening,both relating (22 factors) and not relating (41 factors) to cement, were analyzed on initial postoperative radiographs.Nine THAs with femoral component mechanicalfailure (group 1) were comparedto ( 1) 88 mmfailed THAs (group 2) and (2) 9 matched-paired,nonfailed THAs (group 3). Significant differenceswere evident regardingminimum and maximum cement mantle thickness in Gruen zone 5 and combinedzones5/6 (groups I KS 2 and groups 1 vs 3). with f~ilctl femoralcomponentshaving thinner cementmantles.Discriminateanalysisdetermined minimum cementmantlethicknessin zone 5 to bethe factor mostpredictive of femoral componentfailure. Thesedata indicate that a suboptimal (thin) cement mantle at the medialdiaphysis(Gruen zones5 and 6) contributed to femoralcomponent mechanical looseningin this THA series.This relationship may not pertain to femoral stemsof different materialsor cross-sectionalcharacteristics.Many THA systemscurrently provide for a method of centralization of the femoral component distal stemasa mechanismto ensurean adequatecircumferential distalcementmantle. Continued invotigation into techniques directed toward centralization of the distal femoral stem i,, warranted by the findings of this study. Key words: distal
Total hip arthroplasty (THA) remains the cornerstone of treatment for end-stage coxarthrosis. Total hip arthropiasty has been used to successfully treat many types of hip disease, including osteoarthritis, post-traumatic arthritis, rheumatoid arthritis, inflammatory arthritidcs, congenital dysplasia, meta-

bolic diseases.and developmental disorders, such as Perthes disease and slipped capital femoral epi1(* physis. 5.9.1,.l~,lX.23-25.1x-31.3~ Due to reports of high failure rates of THA ccmented acctabular components, use of noncemented porous-ingrowth acetabular components has become common.‘5.‘7 Early results leave been excelIcnt.‘5.‘” Noncemented femoral components have also been used extensively during the past decade, although early results have not been universally as encouraging.“,” As such, cemented femoral compo-

143

144

The Journal

of Arthroplasty

Vol. 9 No. 2 April

1994

continue to play a ma,jor role in THA, with failure rates declining with the use of modern component designs and cement techniques. 17,3i Continuing analysis of the factors associated with cemented femoral component loosening may provide information on component design and cement technique that could further minimize this cause of failed outcome. Recently, many THA systems have added the optional use of a distal femoral stem ccntralizing mechanism in order to ensure an adequate circumferential distal cement mantle. Although supported by laboratory investigation,“,“,“,” clinical substantiation of the value of a distal stern centralizer has not yet been forthcoming. The purpose of thih study was to evaluate, in a cIinica1 setting, the potcntial benefit of THA femoral component distal stem centralization. A cemented titanium-alloy THA femoral component with a (nonmodular) titanium-bearing surface was retrospectively studied. The early (2-5-year) femoral component failure rate was unacceptably high by modern standards,‘7,‘i largely due 10 the composition and design of the femoral component, I .1.1(1,31 However, this accelerated rate of failure offered an opportune clinical model for the exploration of factors potentially contributing to femoral component failure, and specifically for the exploration of the potential benefit of distal stem ccntralization. llellts

Materials

and Methods

One hundred consecutive cemented THAs, performed between 1982 and 1985, were followed clinically and radiographically for 2-5 years. All THAs were performed by a single senior surgeon (C.W.C.) utilizing a posterior approach. The DF-80 prosthetic system (Zimmer, Warsaw, IN) was used in all cases. The femoral prosthesis was a titanium alloy with a fixed (nonmodular) 28 mm diameter head. Stems were available with diameters of I 3.0, 14.5, 16.0, and 17.5 mm, and were designed to fill 80% of the femoral intramcdullary canal. The canal was prepared using a straight conical reamer and matching broach corresponding to each prosthesis size, as rccommcnded by the manufacturer. Cement technique included pulsatile lavage, canal drying, a distal polyethylene cement restrictor, refrigerated centrifuged polymethyl methacrylate. and pressurized retrograde cement delivery using a cement gun. A femoral stem centralizer was not available with the system. An attempt was made at femoral component insertion to place the stem in neutral or valgus position. Three THAs were excluded from the study, two

with septic femoral loosening and one with aseijtic acetabular loosening without femoral Iooscning. 01 the 97 remaining THAs, 9 failed ftimoral component\ wcrc identified (5 revised for mechanical loosening, 3 with probable radiographic loosening, and I with possibIe loosening’“) and designated asgroup I. The 88 nonfailed femoral components wrrc designatcd as group 2. Group 1 versus Group 2 initial postoperative radiographs (obtained 6 weeks after surgery by a standardired protocol) of the 97 THAs were analyzed regarding femoral coniponent minimum and maximum cement mantle thickness. Measurements were made separately in Grucn zones 1-3 and 5-7, aswell as in various conibinations (zones l/2/3, 213, 5/6/7, 516, and 6/7), for anteroposterior (AP) and lateral radiographic views.“’ Failed (group 1) versus nonfailed (group 2) femoral components were then compared regarding the 22 variables relating to cement mantle thickness. All radiographs were blindly reviewed by a single examiner. Radiographic magnification was standardized by establishing the ratio between the femoral component head size measured on the radiographs and the actual prosthesis head diameter (28 mm). Statistical analysis was performed ~lsing a personal computer software package (NCSS, Kaycvillc, UT). The means for each variable were compared using the t-test for parametric data and Wilcoxon’s test for nonparametric data. Significance was dctcrmined at the conventional !‘< .05 level. The Koilf~rronilDunn modification was applied as warranted for multiple I-test comparisons. Group 1 versus Group 3 A subset of 9 nonfailed lemoral components (group 3) was created from the 88 nonfailed femoral components (group 2). This group 3 subset was selected by malch-pairing with the nine failed femoral components (group I ). The nine tnatched pairs (groups I and 3) wcrc determined on the basis 01 computer-~eneratedierat~d data available for all 100 THAs regarding sex, age, weight, and diagnosis, a5 well a\ activity level as determined by the preoperative Harris hip ccorc. A more extensive radiographic analysis was performed for the nine matched pairs (group I vs group 3). Initial (6-week) postoperative radiographs of groups I and 3 were analyzed, in addition lo cement mantle thickness, for other factors potentially predictive of fkmoral component mechanical loosening. Femoral components were analyzed for center of rotation (in horizontal and vertical planes), offset, limb

Suboptimal

length (operated leg relative to contralateral leg), calcar length, varus/valgus stem alignment, flexion/extension stem alignment, stem tip intramedullary canal centering, and stem to canal ratio. Femoral cement mantles were evaluated separately in Gruen zones 1-7 for cement intrusion, cement voids, intcrface luccncies (bone-cement and prosthesis-cement), and distal cement plug length. Acetabular components were analyzed in the coronal plane for abduction angle and for horizontal and lateral positions relative to the teardrop regarding the potential effects on femoral component loosening. Statistical analysis was as described for the group 1 versus group 2 comparison. However, discriminate analysis was added for the group 1 versus group 3 matched-pair comparison. Discriminate analysis, a statistical technique similar to multiple regression analysis but reserved for the condition in which the dependent variable is discrete rather than continuous, was applied in order to investigate whether femoral component failure could be predicted by any combination of 25 selected variables. Results

Distal Femoral

Cement

Mantle

l

Star et al

145

Fig. 1. Significant differencesbetweenfailed and nonfailed

THAs were found regarding cement rnantlc thickness at the distal medial femoral component (Grucn zones 5 and 6). with failed components having thinner cement thickness.

Group 1 versus Group 2 Comparison of the 9 failed (group 1) to the 88 nonfailed (group 2) femoral components revealed statistically significant differences regarding: ( 1) minimum cement mantle thickness at the medial distal stem in AP Grucn zone 5 (group 1 mean = 0.6 mm, group 2 mean = I .5 mm; P< .0002), (2) maximum cement mantle thickness in AP Gruen zone 5 (group 1 mean = 1.6 mm, group 2 mean = 2.3 mm; P < .004), and (3) minimum cement mantle thickness in AP Grucn zones 5 and 6, when analyzed together (group 1 mean = 1.2 mm, group 2 mean = 1.8 mm; P < ,003) (Fig. 1). Nineteen other variables relating to cement mantle thickness in the AP and lateral zones were not significantIy different (Table 1). In light of the number of variables investigated, the Bonferroni/Dunn modification of the t-test for statistical significance was applied. By considering the 22 variables relating to cement mantle thickness, the Bonferroni/Dunn modification of the t-test would require P < .002 for statistical significance. By this modification, cement mantle thickness differences between failed (group 1) and nonfailed (group 2) femoral components remained significant (AP Gruen zone 5 minimum cement mantle thickness, group 1 thinner; P < .0002) or nearly significant (Gruen zone 5 maximum cement mantle thickness, group 1 thinner; P < ,004 and minimum cement mantle thick-

nessin AP Gruen zones 5 and 6, group 1 thinner; P < .003). Should all 63 variables subsequently studied in group 1 be considered, the Bonferroni/Dunn mo&fication of the t-test would require P < .0008 for statistical significance. By this modification, the most critical finding in the study remained statistically significant. When comparing failed (group 1) versus nonfailed (group 2) THA femoral component distal medial (AP Gruen zone 5) minimum cement mantle thickness, the failed components (group 1) demonstrated a significantly thinner cement mantle (group 1 mean = 0.6 mm, group 2 mean = 1.5 mm; P < .0002).

Group 1 versus Group 3 Comparison of the nine matched pairs (group 1 vs group 3) regarding demographic factors revealed identical sex and diagnosis and no significant differences in age, weight, and activity levels for the groups. This analysis confirmed reasonable selection of matched pairs. Statistically significant differences between failed (group I ) versus nonfailed (group 2) matched pairs were found regarding: (1) minimum cement mantle thickness at the medial distal stem in AP Grucn zone 5 (group 1 mean = 0.6 mm, group 3 mean = 1.7 mm; P < .OI), (2) minimum cement mantle thickness in AP Gruen zones 5 and 6 when analvzed to-

146

The Journal Table

of Arthroplasty

1. Failed

(Group

Vol. 9 No. 2 April 1) versus Nonfailed

1994 (Group

Cernem

mantle mantle mantle

Ninctcen

thickness, thickness, thickness,

other

Table

Grucn Grucn Gruen

variable

2. Failed

with

(Group

5 minimum zont’ 5 maximum zone\ 516 minimum

mnc

mantle mantle

Cement

mantle

SIX& other

thickness. thickness. thickness, variables

Gruen Grucn Gruen with

1) versus Paired

Nonfailed

1 mean

P

I.5

.OOO’

2.3

.004

1.2

1.8

.oo 1

(Group

zone 5 minimum zones 516 minimum zones

516 maximum

3) Femoral

0.6 I.2 7.7

Components:

Significant

Group 3: Nonfailures (11 = 9) (mm)

Differences

-__

1.7 I .‘I 3.9

P .o I .O.! .I)3

I’ 3, .05.

I.2 mm, group 3 mean = 1.9 mm; P < .02), and (3) maximum cement mantle thickness in AP Gruen zones 5 and 6 when analyzed together (group 1 mean = 2.7 mm, group 3 mean = 3.9 mm; P < .04) (Fig. 1). Nineteen other variables relating to cement mantle thickness in the AP and lateral zones, and 60 other variables overall, showed no statistical significance (Table 2). Discriminate analysis was performed for matchpaired groups (group I vs group 3) to determine if any combination of the 25 selected variables could be used to predict femoral component failure. The most predictive factors (highest F value) were: ( 1) minimum cement mantle thickness at the medial distal stem in AP Gruen zone 5 (F = 6.9), (2) maximum cement mande thickness in AP Gruen zone 6 (F = 6.8), and (3) maximum cement mantle thickness in AP Gruen zone 5 (F = 6.7). Twenty variables had an F value under 3.0 (Table 3). Failed (group 1) femoral components had a thinner mean minimum and mean maximum cement mantle thickness than the nonfailed (group 3) femoral components in AP Grucn zones 5 and 6 (Fig. 2). Using only the three most predictive variables listed above, all relating to cement mantle thickness in zones 5 and 6, I5 of 18 group 1 and group 3 femoral components could be correctly predicted as failed or nonfailed (P < ,001). (group

Group 2: Nonfailures (n = 88) (mm)

1 .h

(II = 9) (mm)

Cement

Differences

0.6

Group I: Failure5 Ccmcnt

Significant

P > .05.

Variable

gether

Components:

Group 1: Failures (n = 9) (mm)

Variable Cement Ccmcnt

2) Femoral

=

Discussion The purpose of this study was to evaluate, in a clinical setting, the potential value of centralization of the distal stem regarding THA cemented femoral component mechanical loosening. In order to ac-

complish this evaluation, this THA seric%s was chosen for the following reasons: ( I) our ~nost recent series of ccme~wd THA femoral component\ ( 1989%prescnt), utilizing a modular cobalt-chrome alloy stem with a distal stem centralizer, dots not have sufficient longitudinal follow-up data for meaningful evaluation, (2) our prior series of cemcntcd THA femoral components ( 1985- 1989), utilizing a modular co-

Table 3. Failed (Group 1) versus Paired Nonfailed (Group 3) Femoral Components: Discriminate Analysis F

Value 6.Y

6.X 6.7 5.5 3.3 2 Y 2.7 2.2 I.8 1.7 I .(, 1.3 I.2 I .o 0.8 0.7 0.7 0.7 0.2 0.2 0. I 0. I 0. I 0. I 0.0

Suboptimal

Distal Femoral

Cement

Mantle

l

Star et al.

147

Fig. 2. (A) Initial postoperative radiograph of a bilateral THA. The right femoral component demonstrates an adequate circumferential distal cement mantle. The left femoral compnnent demonstrates a suboptimal (thin) cement mantle along the distal medial stem, in Gruen zone 5. (arrows). (B) Sixteen-month postoperative radiograph of the same bilateral THA. The left femoral component, with a suboptimal cement mantle noted on the initial postoperative radiograph, has failed due to aseptic loosening (subsidence constituting definite loosening’“). The right femoral component, with an adequate distal cement mantle, remains stable.

balt-chrome alloy stem without a distal stem centralizer, has not (fortunately) produced a sufficient number of failed femoral components for meaningful evaluation, and (3) this series( 1982- 1985) provided a sufficient number of femoral component failures to attempt evaluation of the potential benefits of distal stem centralization. Although the great majority (9 1 of 100) of femoral components in this THA series remained mechanically stable during the follow-up period, femoral component failure occurred more frequently in this series than would be expected with modern cementing techniques.‘7.33 Factors implicated for the high failure rate of this particular prosthesis include the titanium-bearing surface of the fixed head and the design of the reamers and component stem.‘.‘, 10,‘7 Yet, because the relatively large number of failures (9 of 100) allowed comparison with nonfailure and match-paired nonfailure groups, this series offered an opportune model for the evaluation of the potential benefit of femoral component distal stem centralization. The results of this study indicate that femoral component mechanical loosening in this THA series was associated with a suboptimal, thin cement mantle along the distal medial aspect of the femoral component stem in Gruen zones 5 and 6. These data indiloosening would have been decate that mechanical

creased by achieving a thicker, more optimal cement mantle along the medial diaphysis, such as would have been achieved by reproducible centralization of the distal stem. There were no significant differences between the failed and nonfailed femoral components among the other (more than 60) variables studied. This occurrence does not allow for conclusions regarding the relevance of these other factors to femoral implant outcome. Rather, it indicates that, whether relevant or irrelevant to implant outcome, these factors were treated equally by a single surgeon utilizing a single surgical approach and a single type of prosthesis in this THA series. Historically, varus positioning of the femoral component has been implicated as contributory to mechanical loosening due to increased bending moment and decreased axial loading on the stern. ’ Many authors have therefore recommended neutral or valgus positioning of the stem.6 In this series, however, our data indicate that excessive valgus positioning, when associated with medial distal stem placement compromising optimal thickness of the distal medial cement mantle (Gruen zones 5 and 6), was associated with mechanical loosening. While actual varus/valgus angle measurements had no significant effect, absolute measurements of cement mantle thickness along the medial diaphysis revealed a highly signifi-

148

The Journal

of Arthroplasty

Vol. 9 No. 2 April

1994

cant difference between failed and nonfailed groups, with the successful femoral components having a greater mean cement mantle thickness in these zones. This would indicate that excessive valgus positioning, when associated with excessive medial displacement of the distal stem and a suboptimal cement mantle in Gruen zones 5 and 6, contributes to mechanical loosening and femoral component failure. Improved centralization of the femoral component distal stem could be assumedto have, in turn, optimized distal cement mantle thickness in this THA series.Eglund et al. showed that use of a finned distal centralizing device improved centralization in both mediolateral and AP planes.’ Though the application of such a centralizing device has not yet been shown to positively influence outcome in a clinical THA series, its benefit is implied by the data in this study. However, this inverse relationship between distal ccment mantle thickness and femoral component failure may not be applicable to femoral components with different physical properties and configurations. Prior finite element analysis, bench cadaver, and autopsy retrieval studies have demonstrated a high stressand strain concentration at the cemented femoral component distal cement mantle. Further, axial compression strain is increased (on finite element analysis) by a thinner cement mantle.’ Harrigan and Harris”,‘3 studied stressdistribution in the cement mantle of a cemented femoral component with finite element analysis. They showed that: (1) the most likely region for initial failure is the cement-metal interface and (2) stressesat the medial margin of the distal stem tip are sufficient to result in cement-metal interface disruption (debonding). Further, before and after interface debonding, peak stressesat the distal stem tip are essentially equal; but after interface debonding, the most significant increase in stresswithin the distal cement mantle occurs medially near the stem tip. Estok et al.x studied cement strains near the tip of a cemented femoral component with finite clement analysis. They showed that: ( 1) peak axial comprcssion strain within the entire cement mantle occur’r medially near the tip of the stem, (2) with an increase in cement mantle thickness from 2.5 to 5.0 mm, a marked reduction in cement mantle strain occurs, and (3) with that increase in cement mantle thickness,peak axial compression strain within the entire cement mantle remains medially near the tip of the stem but decreasesby 46%. O’Connor et al.” showed compressive strain magnitude at the medial tip of cemented femoral components in cadaver human fcmora, with simulated single leg stance and stairclimbing to be nearly twice

that of the mean magnitude throughout the cement mantle. They noted that increased strain at the distal stem implicated this region as likely for initiating fracture of cement fixation. They further correlated these bench findings to autopsy studies of successful long-term cemented THAs that showed increased bone strain and cement-metal interface disruption at the level of the femoral component stem tip.‘y,ZZ This clinical study indicates that a suboptimal distal cement mantle is associated with early femoral component failure. Radiographic data correlate a thin femoral mantle at the distal medial femoral stem with femoral component loosening. This early mechanical failure, taken in light of the studies outlined above, suggestsa mechanism of cement fragmentation or fracture in regions of thin cement and high stress and strain at the medial stem tip. In conclusion, mechanical loosening of cemented femoral components in this THA series was significantly associatedwith a thinner cement mantle along the distal medial aspect of the stem in Gruen zmcs 5 and 6. Although factors other than cement mantle thickness contributed to early failure, the unacceptably high femoral component mechanical loosening rate in this THA series would have been decreased by achieving a more optimal distal cement mantlc. Many THA systems currently provide the means for distal stem centralization, designed to ensure an adecluate circumferential distal cement mantle. Continued investigation into the techniques of cemented femoral component distal stem centralization is warranted by the findings of this study.

References I. Agins HJ, Alcock NW, Bansal M el ~1: Metallic wc’dr in failed titanium-alloy total hip replacements: a histological and quantitalivc analysi\. .I Bone Joint Surg 70A: 347, 1988 2. Black .I, Sherk H, Bvnini J et ~1: Mct;lllo3is a\sotiateci with 2 stable titanium-alloy femoral compont’nl in total hip rcplaccment. .I Boric Joint Suq 72A: 126. 1990 3. CalandruccioRA: Arthroplasty. p, 2 192. In Etimonson AS, Crenshaw AH (eds): Campbell’s operative orthopaedics. 6th ed. CV Mosby, Sr. Loui,, 1980 4. Callaghan JJ, Dysart SH, Savory CG: The uncementtd porous-coated anatomic total hip prosrhesi\. J Bone Joint Surg 7OA:337, 1988 5. Clark HJ: Total replacement ol rhe hip lot- avascular necrosis in sickle cell disease. J Bow Joint Surg 7 I B: 465, 1989 6. Eftekhar NS: Principles of tolal hip arthroplasty. CV Mosby, St. Louis, 1978 7. Eglund N, Lidgen L, Onnerfalt R: Improved position-

Suboptimal

8.

Y.

IO.

11. 12.

13.

14.

15. 16.

17.

18.

19.

20.

21.

ing of the femoral stem with a centralizing device. Acta Orthop Stand 61:236, 1990 Estok DM, Harrigan TP, Harris WH: Finite element analysis of cement strains at the tip of an idealized cemented femoral component. Trans Orthop Res Sot 16:504, 1991 Gitclic S: The treatment of pigmented villonodular synovitis of the hip: a case report and literature review. Clin Orthop 239: 154, 1989 Gruen TA, McNiece GM, Amstutz HC: ‘Modes of failure’ of cementecl stem-type femoral components: a radiographic analysis of loosening. Clin Orthop 14 1 : 17, 197Y Harold HZ: Congenital dislocation of the hip treated by total hip arthroplasty. Clin Orthop 242: 195, 1990 Harrisan TP, Harris WH: Mechanical consequences of cement-metal interface disruption (debonding) at the distal tip of l’emoral total hip prosthesis. Trans Orthop Res Sot 16:503, 1991 Harrigan TP, Kareh J, Harris WH: Initial loosening mechanisms in cemented femoral stems: debonding at the cement-metal interface. Trans Orthop Res Sot 16:520. 1991 Harris CM, Baum J: Involvement of the hip in juvenile rheumatoid arthritis: a longitudinal study. J Bone Joint Surg 70A:82 1, 1988 Harri\ WH, Maloney WJ: Hybrid total hip arthroplasty. Clin Orthop 249:2 I, 1989 Harri\ WH, McCarthy JC Jr, O’Neill DA: Femoral co~~qmw~t loosening using contemporary techniques of femoral c‘cment fixation. J Bone Joint Surg 64A: 1063, lY82 Harris WH, McCann WA. Loosening of the femoral component after use of the medullary-plug cementing technique. J Bone Joint Surg 68A: 1064, 1986 Hartofilakidis G: Low friction arthroplasty for old untreated congenital dislocation of the hip. J Bone Joint Surg 7Of3: 182, 1988 .Jabty M. Maloney WJ, Bragdon CR et al: Histomorphological studies of the long-term skeletal responses to well-fixed cemented femoral components. J Bone Joint Surg 77A: 1220, 100 Lombardi AV, Mallory TH, Vaughn BK, Drouillard P: Aseptic loosening in total hip arthroplasty secondary to osteolysi\ induced by wear debris from titanium alloy modular femoral heads. J Bone Joint Surg 7 1 A: 1337, lY8Y Malonqz WJ, Harris WH: Comparison of a hybrid with an uncerncmted total hip replacement. J Bone Joint Surg 77A: 1349, 1990

Distal Femoral

Cement

Mantle

l

Star et al.

149

22. Maloney WJ, Jasty M, Burke DW et al: Biochemical and histologic investigation of cemented total hip arthroplasties. Clin Orthop 24Y: 12Y, 1YXY MP, Weinstein SL: A long-term follow-up 23. McAndrew of Legg-Calve-Perthes disease. .J Boric Joint Surg 66A: 860, 1984 24. McQuery FG, Johnston RC: Coxarthrosis after congenital dysplasia: treatment by total hip arthroplasty without acetabular bone grafting. .I Boric Joint Surg 70A:ll40, 1988 25. Merkow RL, Pellicci PM, Hely DP, Salvati EA: Total hip rcplacerncnt for Paget’s disease of the hip. .I Bone Joint Surg 66A:752, 1984 26. Morscher E, Bereiter H, Lampert C Cementles\ pressfit cup. Clin Orthop 249: 12, 1989 27. O’Comior D, Burke DW, Scdlacek KC. Harris WH: Peak cement strains in cemented femoral total hip. Tram Orthop Res Sot 16:220, 1YY 1 28. Pass R, Malone JP, Ewald FC et al: Six-to-clevcn-year results of total hip arthroplasty in rheumatoid arthritis. Clin Orthop 182: 109, 1984 29. Radford PJ, Doran A, Greatorex RA. Rushton N: Total hip replacement in the renal transplant recipient. J Bone Joint Surg 7 1B:456, 198Y 30. Ranawat CS, Atkinson RE, Salvati EA. Wilson PD Jr: Conventional total hip arthroplasty for degenerative joint disease in patients between the ages of 40 and 60 years. J Bone Joint Surg 66A:745, 1984 31. Ranawat CS, Dorr LD, lnglis AE: Total hip arthroplasty in protrusio acetabuli of rheumatoid arthritis, J Bone Joint Surg 62A: 1059, 1980 32. Robinson RP, Love11 TP, Green TM, Bailey GA: Early fernoral component loosening in DF-80 total hip arthroplasty. J Arthroplasty 4:55, 1989 33. Russotti GM, Coventry MB, Stauffer RN: Cemented total hip arthroplasty with contemporary techniques. Clin Orthop 235: 141. 1988 34. Salvali EA. Wilson PD Jr, Jolley MN ct al: A ten-year follow-up study of our first 100 consecutive Charnley total hip replacements. J Bone .Joint Surg 63A:753, 1981 35. Stauffer RN: Ten-year follow-up stucl~~ 01’ total hip replacement. J Bone Joint Surg 64A:983, 1982 36. Strathy GM, Fitzgerald RfI: Total hip ar-throplasty in the ankylosed hip. .J Bone .Joint Sllrg 7OA:96 3, 1988 37. Sutherland CJ, Wilde AH, Borden LS, Marks ICE: A ten-year follow-up of 100 con\eculivc’ Muller curvedstem total hip-replacement arthroplasly. J Bone Joint Sur: 64A:Y70. 1988