JBJS | The Accuracy and Reproducibility of Radiographic Assessment of Stress-Shielding : A Postmortem Analysis*

The Journal of Bone & Joint Surgery, Volume 82, Issue 10

Articles | October 01, 2000

The Accuracy and Reproducibility of Radiographic Assessment of Stress-Shielding : A Postmortem Analysis*

C. Anderson EnghJr., M.D.†; James P. McAuley, M.D.†; Christi J. Sychterz, M.S.†; Marie E. Sacco, B.S.†; Charles A. Engh, Sr., M.D.†

View Disclosures and Other Information

Investigation performed at the Anderson Orthopaedic Research Institute, Alexandria, Virginia
*One or more of the authors has received or will receive benefits for personal or professional use from a commercial party related directly or indirectly to the subject of this article. No funds were received in support of this study.
†Anderson Orthopaedic Research Institute, 2501 Parkers Lane, Suite 200, Alexandria, Virginia 22306.

The Journal of Bone & Joint Surgery. 2000; 82:1414-1414

5 Recommendations (Recommend) | 3 Comments | Saved by 3 Users Save Case

text A A A

Abstract

Background: Although periprosthetic bone loss remains a major concern in total hip arthroplasty, radiographic assessment of such loss is both difficult and subjective. In the present study, we assessed the ability of orthopaedic surgeons to reproducibly recognize changes in periprosthetic bone density on radiographs. We hypothesized that assessment of periprosthetic bone loss on plain radiographs is not reliable enough to justify its use in outcomes research.

Methods: Twenty-nine unilateral total hip replacements and the surrounding bone were retrieved at autopsy, and radiographs were made; radiographs of the contralateral, normal femur were also made after implantation of an identical prosthesis and used as a control. Three orthopaedic surgeons independently examined the specimen radiographs and classified bone loss in each of sixteen femoral zones. Bone loss was recorded as present if the bone of the femur that had had in vivo implantation showed evidence of cortical thinning, increased porosity, or decreased density (either cortical or trabecular) when compared with the control femur. The kappa coefficient was used to quantify interobserver and intraobserver reproducibility in determining bone loss for the 464 zones examined and in determining the Engh and Bobyn stress-shielding classification of each femur. In fourteen femoral pairs, bone loss was also quantified with dual-energy x-ray absorptiometry, and the resulting value was then compared with the bone-loss classification that had been determined radiographically.

Results: First, the surgeons agreed on the presence or absence of bone loss in 73 percent (337) of the 464 zones. The interobserver kappa value of 0.58 denoted only good reproducibility. The intraobserver reproducibility was better; the surgeon's initial evaluation of bone loss agreed with his second evaluation for 90 percent of the zones (kappa = 0.74). Second, the three surgeons agreed on the degree of stress-shielding, according to the Engh and Bobyn classification, in 66 percent (nineteen) of the twenty-nine femora. The kappa value for this comparison was only 0.27, indicating marginal reproducibility. Third, although there was some agreement among reviewers when there was 20 to 60 percent reduction in bone-mineral content as determined with dual-energy x-ray absorptiometry, excellent agreement among the examiners (kappa = 0.85) was not achieved until bone loss averaged 70 percent.

Conclusions: On the basis of these results, we suggest caution in interpreting results from studies of femoral bone loss that have used plain radiographic analysis if the authors have not provided interobserver reliability data. We question the utility of evaluating periprosthetic bone loss on radiographs, since the loss is not reproducibly recognized until 70 percent of the bone is gone.

Figures in this Article

Introduction

Introduction | Materials and Methods | Results | Discussion | References

Periprosthetic bone loss remains a major concern in total hip arthroplasty. Bone loss presents as cortical thinning, decreased cortical density, or decreased trabecular density on radiographs. Although dual-energy x-ray absorptiometry and videodensitometry measure periprosthetic femoral bone loss more accurately^7,8,12,17-19, surgeons rely on radiographs for such evaluations^{2,4-6,9,10,13,16,20-22,24,26,29,32}.

Radiographic assessment of bone loss has limitations that compromise its utility for clinical research. First, assessment depends upon consistent radiographic technique for serial radiographs. Even slight differences in radiographic penetration or patient position affect the ability to detect bone resorption accurately on serial radiographs³¹. Second, visual assessment of cortical thinning and decreased bone density are subjective and, therefore, prone to interobserver error; this makes comparison of results from different observers at different institutions difficult. Finally, we do not know how much bone must be lost around a hip replacement before the loss becomes visible radiographically.

In the present study, we assessed the ability of surgeons to recognize changes in periprosthetic bone density on radiographs. Using femora retrieved at autopsy, we attempted to create ideal radiographic circumstances under which three specialized joint-replacement surgeons from the same institution could evaluate periprosthetic bone loss. We used the sixteen femoral periprosthetic zones defined by Engh and Bobyn for the semiquantitative analysis of stress-shielding⁶. We chose this technique because multiple authors and institutions have used it to classify stress-shielding occurring with total hip arthroplasty^{2,13,16,20,21,24,26,29,32}. We assessed interobserver reproducibility and correlated bone loss determined radiographically with bone loss determined with dual-energy x-ray absorptiometry. We hypothesized that assessment of periprosthetic bone loss on plain radiographs is not reliable enough to justify its use in outcomes research.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 1:: Example of anteroposterior radiograph used to assess periprosthetic bone loss. Bone loss was considered present if the bone of the femur in which the implant had been placed in vivo showed evidence of cortical thinning, increased porosity, or decreased density when compared with the control femur. By making a single radiograph of each matched pair of femora, we ensured identical radiographic conditions for comparisons of bone density. Moreover, by using femora retrieved at autopsy, we sought to eliminate soft tissue and, thus, to optimize radiographs for visualization of bone.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2:: Illustration showing the mean bone loss from fourteen pairs of femora, as determined with dual-energy x-ray absorptiometry in each of the sixteen zones examined.

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 3:: Zones were sorted into groups of thirty-two according to the amount of bone loss (in ascending order) in each. This illustration shows the mean amount of bone loss for each group, as determined with dual-energy x-ray absorptiometry (DEXA), graphed against the percentage of cases in which the doctors agreed or disagreed on the presence of bone loss.

View Larger

Table I: Grouped Dual-Energy X-Ray Absorptiometry Results

Group	Mean Bone Loss (percent)	Agreement (percent)			Kappa Coefficient	P Value
Bone Loss Not Present	Bone Loss Present	Total
1 (n = 32)	-14.9	68.8	3.1	71.9	0.32	0.12
2 (n = 32)	3.3	25.0	37.5	62.5	0.49	<0.01
3 (n = 32)	17.3	25.0	31.3	56.3	0.41	<0.01
4 (n = 32)	31.3	6.3	53.1	59.4	0.30	0.05
5 (n = 32)	50.1	0.0	75.0	75.0	0.11	0.71
6 (n = 32)	69.9	6.3	90.6	96.9	0.85	0.02
7 (n = 32)	96.3	6.3	87.5	93.8	0.73	0.02

Add to My JBJS

Table I: Grouped Dual-Energy X-Ray Absorptiometry Results

Group	Mean Bone Loss (percent)	Agreement (percent)			Kappa Coefficient	P Value
Bone Loss Not Present	Bone Loss Present	Total
1 (n = 32)	-14.9	68.8	3.1	71.9	0.32	0.12
2 (n = 32)	3.3	25.0	37.5	62.5	0.49	<0.01
3 (n = 32)	17.3	25.0	31.3	56.3	0.41	<0.01
4 (n = 32)	31.3	6.3	53.1	59.4	0.30	0.05
5 (n = 32)	50.1	0.0	75.0	75.0	0.11	0.71
6 (n = 32)	69.9	6.3	90.6	96.9	0.85	0.02
7 (n = 32)	96.3	6.3	87.5	93.8	0.73	0.02

Materials and Methods

Introduction | Materials and Methods | Results | Discussion | References

Twenty-nine porous-coated Anatomic Medullary Locking femoral components (AML; DePuy, Warsaw, Indiana) and the surrounding bone were retrieved at autopsy from twenty-nine patients. Seventeen patients were male, and twelve were female. The mean age at the time of the hip replacement was 73.5 years (range, 54.5 to eighty-seven years), and the components were in situ for a mean of 80.5 months (range, seventeen to 162 months). All implants were stable at the time of death.

At autopsy, we retrieved both the femur in which the prosthesis had been implanted in vivo and the contralateral, normal femur. All soft tissue was removed, and the femora were fixed in a 10 percent buffered formalin solution. Subsequently, a prosthesis that was identical in size and extent of porous coating to the prosthesis inserted in vivo was implanted in the contralateral femur. The contralateral femur then served as a control.

Radiographic Evaluation

To assess periprosthetic bone loss, we made anteroposterior and lateral radiographs of each pair of femora, using femoral anatomy to consistently orient the pairs. By making a single radiograph of each pair of femora, we ensured identical radiographic conditions for comparisons of bone density (Fig. 1). Moreover, by using femora retrieved at autopsy, we sought to eliminate soft tissue and, thus, to optimize radiographs for visualization of bone.

Using the technique of Engh and Bobyn⁶, three orthopaedic surgeons independently examined these radiographs for evidence of bone loss. During the examinations, the surgeons had no knowledge of the other surgeons' results or of the dual-energy x-ray absorptiometry data. For the radiographic analysis, the anteroposterior and lateral images of each femur were divided into four equal periprosthetic levels (Fig. 2). These four levels were further divided into medial, lateral, anterior, and posterior zones, yielding a total of sixteen zones for the examination of each femur. In each zone, the bone of the femur in which the implant had been placed in vivo was compared with the bone of the control femur. Bone loss was considered present if the bone of the femur with in vivo implantation showed evidence of cortical thinning, increased porosity, or decreased density (either cortical or trabecular). Bone loss in each zone was simply recorded as present or absent; no attempt was made to quantify the amount of loss radiographically. A total of 464 zones were examined for evidence of bone resorption: eight zones on the anteroposterior radiograph and eight zones on the lateral radiograph were evaluated for each of the twenty-nine femora with in vivo implantation.

With use of the technique of Engh and Bobyn⁶, the degree of stress-shielding for each femur was classified according to the number of zones (of a possible sixteen) that had bone loss. First-degree (mild) stress-shielding indicated one to four sites of bone resorption; second-degree (moderate) stress-shielding, five to seven sites of resorption; and third-degree (severe) stress-shielding, eight or more sites of resorption. As previously stated, this semiquantitative method of assessing stress-shielding was chosen because it has been used extensively by multiple authors and institutions to classify bone loss associated with total hip arthroplasty^{2,13,16,20,21,24,26,29,32}.

We used the kappa coefficient to quantify the extent of agreement among the three surgeons for the 464 zones examined and for the Engh and Bobyn classification⁶ of stress-shielding for each of the twenty-nine femora. The extent of agreement about whether a zone had bone loss or not reflects the reproducibility (or the interobserver reliability) of the measurement¹¹. The kappa coefficient corrects for the chance agreement that would occur if the measurements were totally unrelated. When measurements agree only at a chance level, the value of the kappa coefficient is zero. When observations agree perfectly, the value of the kappa coefficient is one. We considered kappa values from 0 to 0.4 to denote marginal reproducibility²⁵, values of greater than 0.4 and less than 0.75 to denote good reproducibility, and values of 0.75 and greater to denote excellent reproducibility^25,28. Although these designations of marginal, good, and excellent reproducibility are somewhat arbitrary, they are historically accepted and have been quoted frequently in the orthopaedic literature^3,15,27,30. SPSS statistical software (version 8.0; SPSS, Chicago, Illinois) was used to determine kappa coefficients for bone-resorption measurements among the three observers. For maximum statistical power, we treated the 464 zones as independent measures. We acknowledge the possibility that these measures were not truly independent because a surgeon's assessment of bone loss in one zone could have been influenced by his assessment in other zones on the same radiograph. Therefore, our kappa value may be inflated. As such, these results represent the highest possible measure of agreement for assessing bone loss.

To assess intraobserver reproducibility, one surgeon reevaluated the bone resorption in the 464 zones two weeks after the initial evaluation. The surgeon had no record of the results from his initial evaluation or from the dual-energy x-ray absorptiometry analysis. The kappa coefficient was used to compare agreement between this surgeon's first and second evaluations.

Dual-Energy X-Ray Absorptiometry Analysis

The bone-mineral density in fourteen of the twenty-nine femoral pairs was quantified with dual-energy x-ray absorptiometry (DPX; Lunar Radiation, Madison, Wisconsin) according to a previously published protocol⁷. Specimens were scanned in both the anteroposterior and the lateral position, with the implant's shape used to guide femoral orientation. In both views, the dual-energy x-ray absorptiometry software was used to determine bone-mineral density (grams per square centimeter) in the same zones that were studied radiographically. For each zone, bone resorption due to remodeling was calculated as the difference in bone-mineral density between the femur in which the implant had been placed in vivo and the corresponding zone of the control femur. The amount of bone loss determined with dual-energy x-ray absorptiometry was then compared with the bone-resorption classification determined radiographically. The remaining fifteen femoral pairs were unavailable for this analysis because they had been destructively tested in another study.

For graphical presentation of the dual-energy x-ray absorptiometry and reproducibility data, we sorted zones into groups of thirty-two according to the amount of bone loss in each. The thirty-two zones with the least amount of bone loss were placed in Group 1, the thirty-two zones with the next smallest amount of bone loss were placed in Group 2, and so on. Next, we calculated the mean amount of bone loss for each group and graphed this against the percentage of cases in which the doctors agreed or disagreed on the presence of bone loss (Fig. 3).

Results

Introduction | Materials and Methods | Results | Discussion | References

Radiographic Analysis

The three surgeons agreed on the presence or absence of bone loss in 73 percent (337) of the 464 zones that they examined. The kappa value for this comparison was 0.58, which denoted good interobserver reproducibility²⁵. The zones with the highest degree of agreement were the most distal medial zone (kappa = 0.73) and the most distal posterior zone (kappa = 0.62). The surgeons appeared to have the greatest difficulty in assessing bone resorption in the most proximal lateral zone; they agreed on only 45 percent of these zones (kappa = 0.24).

Bone resorption was noted in a mean of 10.9 of the sixteen zones per femur. When classifying the degree of bone resorption as first, second, or third on the basis of the number of zones with bone loss per femur, the three surgeons agreed on 66 percent (nineteen) of the twenty-nine cases. These nineteen femora were classified as having third-degree bone resorption. The kappa value for this comparison was 0.27, indicating only marginal reproducibility. The smaller kappa value was due in part to the small number of cases (twenty-nine) for this comparison, as opposed to the large number of observations (464) for the first comparison.

When one surgeon assessed bone resorption at two different times, his initial evaluation of bone loss agreed with his second evaluation for 90 percent of the zones. The kappa value for this comparison was 0.74 and denoted nearly excellent intraobserver reproducibility.

Dual-Energy X-Ray Absorptiometry Analysis

Percent changes in bone-mineral density between the femur in which the implant had been placed in vivo and the control femur were calculated for 224 zones (fourteen femoral pairs multiplied by sixteen zones per pair). The mean amount of bone loss (and standard deviation) per zone was 38.1 ± 35.8 percent and ranged from 8.5 to 67.7 percent (Fig. 2).

We then determined whether the amount of bone loss measured by dual-energy x-ray absorptiometry was related to the ability of the surgeons to detect it radiographically and to agree upon its presence or absence. Bone loss averaged 58.4 ± 28.8 percent in the zones in which all surgeons agreed that bone resorption was present. Bone loss averaged 22.8 ± 23.7 percent in the zones in which the surgeons disagreed on the presence or absence of bone loss. Bone loss averaged 4.6 ± 31.0 percent in the zones in which all surgeons agreed that there was no bone loss.

As described, the mean amount of bone loss was calculated for each group of thirty-two zones and plotted against the percentage of cases in which the doctors agreed or disagreed on the presence of bone loss (Fig. 3). The line indicating the percentage of cases in which the surgeons agreed upon the presence of bone resorption had a rising slope; this revealed that the surgeons were better able to detect and agree upon radiographic bone loss as the amount of that loss increased. In the curve that indicated the percentage of cases in which the surgeons disagreed upon the presence of bone resorption, the peak demonstrated that the greatest uncertainty in determining bone loss occurred when there was between 10 and 30 percent loss. Moreover, this curve demonstrated a substantial amount of disagreement (more than 20 percent) until bone loss reached 50 percent. The kappa value, calculated for each of the seven groups on the curve, did not demonstrate excellent interobserver reproducibility (kappa = 0.85) until a mean of 70 percent of the bone had been lost (Table I).

Discussion

Introduction | Materials and Methods | Results | Discussion | References

Adaptive changes in the femur after total hip arthroplasty usually are assessed on plain radiographs with little information on the reliability or accuracy of the assessment. The purpose of this study was to evaluate the ability of surgeons to reproducibly recognize periprosthetic bone loss on plain radiographs. Using autopsy-retrieved femora that were free of soft tissue and identically positioned, with each pair of femora included on the same radiograph, we attempted to create ideal radiographic conditions. We believed that if three surgeons did not agree on the presence or absence of bone resorption on these radiographs, then conventional serial radiographic measurements of bone loss would be called into question. Additionally, dual-energy x-ray absorptiometry provided precise quantification of bone loss and a basis for comparison with the surgeons' radiographic assessments. This allowed us to determine how much periprosthetic bone-mineral loss generally occurred before the surgeons could reliably agree upon the presence of bone loss.

There were three major findings of this study. First, despite having ideal radiographs, the surgeons agreed on the presence or absence of bone loss in only 73 percent (337) of the 464 zones. The interobserver kappa value of 0.58 denoted only good reproducibility. The intraobserver reproducibility was better. The surgeon's initial evaluation of bone loss agreed with his second evaluation for 90 percent of the zones (kappa = 0.74). Second, the three surgeons agreed on the classification, with use of the system of Engh and Bobyn⁶, of the degree of stress-shielding in 66 percent (nineteen) of the twenty-nine femora. The kappa value for this comparison was only 0.27, indicating marginal reproducibility. Third, although there was some agreement among reviewers when there was a 20 to 60 percent reduction in bone-mineral content (Fig. 3), excellent agreement (kappa = 0.85) among examiners was not achieved until a mean of 70 per cent of the bone had been lost (Table I).

Many articles have been written on the accuracy and the intraobserver and interobserver reliability, as indicated by kappa values, of various orthopaedic classification systems^3,15,27,30. Kappa values between 0.4 and 0.75 have been accepted as denoting good reproducibility and values greater than 0.75, excellent reproducibility^25,28. Generally, it is unsound to compare studies unless there is excellent interobserver reproducibility for the classification system or the observation being reported on. There are several reasons why a classification system or an observation might not have excellent reproducibility. One common reason is a poorly defined or overly complex classification system. We do not believe that this was the case in our study, since the reviewers only had to determine if a patient had either a thinner cortex or less bone density in a well defined zone. Another reason for poor reproducibility is variation in radiographic technique. It can be difficult to tell if an apparent change in bone density represents an actual loss of bone or simply a different radiographic setting³¹. Even with identical settings, different film and processing chemicals used on different days can affect the appearance of bone density on serial radiographs. Finally, identical rotation is difficult to obtain with serial radiographs of the same patient. Because we made radiographs of postmortem specimens, we were able to control for these variables better than is possible in a clinical setting, and we were able to create optimum radiographs for visualization of changes in bone density. However, despite a straightforward definition of bone loss and the creation of optimum radiographs, the kappa value for interobserver reproducibility in our study was only 0.58 (good), and the three observers agreed upon bone loss in only 73 percent of the zones. Although the intraobserver variability approached the excellent category (kappa = 0.74), we do not know whether this result can be duplicated with use of serial clinical radiographs. On the basis of these results, we question the ability of physicians to reproducibly evaluate stress-shielding on serial radiographs of patients.

With regard to the second finding, many authors have recognized the difficulty involved in the radiographic classification of periprosthetic femoral bone loss. The system of Engh and Bobyn is somewhat quantitative since it is based on the total number of zones in which bone loss is observed on serial radiographs⁶. This classification has been used by multiple authors to report periprosthetic bone loss associated with cementless, cemented, primary, and revision total hip replacements^{2,13,16,20,21,24,26,29,32}. It is intuitive that evaluation of multiple periprosthetic zones, rather than a single periprosthetic zone, could lead to a more accurate classification of bone loss. However, the reproducibility of this classification system has not been tested previously to our knowledge. Our reviewers used the classification system for twenty-nine femora and agreed on the degree of stress-shielding in 66 percent (nineteen) of them. The poor interobserver kappa value (0.27) indicated that the system is not reproducible. However, it is possible that the poor reproducibility in this study was due to the small number of femora (twenty-nine) that were examined and the fact that most were classified as having third-degree stress-shielding.

Our final finding was based on comparison of radiographic observations with dual-energy x-ray absorptiometry data. Dual-energy x-ray absorptiometry allowed us to determine how much bone-mineral content was lost before the observers agreed that less bone was seen on the radiographs. The radiographic literature has stated that greater than 30 percent of bone must be lost before it can be detected on radiographs^1,14,23. However, these studies did not involve periprosthetic bone loss. More pertinent to our study, McGovern et al. reported that radiodensitometry had confirmed that periprosthetic bone loss between 15 and 36 percent was visible on plain radiographs¹⁸. However, this statement was based on the analysis of only three postmortem specimens. In another study, Kilgus et al. stated that changes in bone density were obvious radiographically even though the greatest decreases in bone-mineral density measured with dual-energy x-ray absorptiometry were on the order of 35 percent¹². However, Kilgus et al. did not define how bone loss was measured radiographically and made no reference to the intraobserver variability. In our study, when bone loss measured by dual-energy x-ray absorptiometry averaged 31 and 50 percent, the three observers agreed on the presence of bone loss in only 53 and 75 percent of the zones, respectively (Table I). It was not until bone loss averaged 70 percent that the reviewers agreed upon the presence of bone loss in more than 90 percent of the zones and the kappa value was excellent (0.85).

In this study, we attempted to create a best-case scenario for testing the reproducibility of radiographic assessment of stress-shielding. We made radiographs of twenty-nine pairs of autopsy-retrieved femora under ideal conditions and had three specialized joint-replacement surgeons review those radiographs. We chose to treat the 464 zones (sixteen zones for each of the twenty-nine femora) as independent measures in order to maximize statistical power. Although the surgeons attempted to analyze each zone independently, we acknowledge the possibility that the result for one zone on a radiograph could have influenced analysis of adjacent zones on that radiograph. If that was the case, the zones were not truly independent measures and our kappa value is inflated. As such, our results represent the highest possible measure of agreement for the assessment of bone loss radiographically. Even so, we failed to find excellent interobserver reproducibility. On the basis of these results, we question the ability of physicians to reproducibly evaluate stress-shielding on serial radiographs of patients and we suggest caution in interpreting studies of periprosthetic femoral bone loss on plain radiographs if the studies do not provide interobserver data. Finally, we question the utility of evaluating periprosthetic bone loss on radiographs since, even under ideal conditions, bone loss is not reproducibly recognized until 70 percent of the bone is gone.

References

Introduction | Materials and Methods | Results | Discussion | References

Ardran, G. M.: Bone destruction not demonstrable by radiography. British J. Radiol., 24: 107-109, 1951.

Bugbee, W. D.; Culpepper, W. J., II; Engh, C. A., Jr.; and Engh, C. A., Sr.: Long-term clinical consequences of stress-shielding after total hip arthroplasty without cement. J. Bone and Joint Surg., 79-A: 1007-1012, July 1997.

Burstein, A. H.: Editorial. Fracture classification systems: do they work and are they useful? J. Bone and Joint Surg., 75-A: 1743-1744, Dec. 1993.

Cohen, B., and Rushton, N.: Bone remodelling in the proximal femur after Charnley total hip arthroplasty. J. Bone and Joint Surg., 77-B(5): 815-819, 1995.

D'Antonio, J. A.; Capello, W. N.; and Manley, M. T.: Remodeling of bone around hydroxyapatite-coated femoral stems. J. Bone and Joint Surg., 78-A: 1226-1234, Aug. 1996.

Engh, C. A., and Bobyn, J. D.: The influence of stem size and extent of porous coating on femoral bone resorption after primary cementless hip arthroplasty. Clin. Orthop., 231: 7-28, 1988.

Engh, C. A.; McGovern, T. F.; Bobyn, J. D.; and Harris, W. H.: A quantitative evaluation of periprosthetic bone-remodeling after cementless total hip arthroplasty. J. Bone and Joint Surg., 74-A: 1009-1020, Aug. 1992.

Engh, C. A.; O'Connor, D.; Jasty, M.; McGovern, T. F.; Bobyn, J. D.; and Harris, W. H.: Quantification of implant micromotion, strain shielding, and bone resorption with porous-coated anatomic medullary locking femoral prostheses. Clin. Orthop., 285: 13-29, 1992.

Heekin, R. D.; Callaghan, J. J.; Hopkinson, W. J.; Savory, C. G.; and Xenos, J. S.: The porous-coated anatomic total hip prosthesis, inserted without cement. Results after five to seven years in a prospective study. J. Bone and Joint Surg., 75-A: 77-91, Jan. 1993.

Jacobs, J. J.; Sumner, D. R.; and Galante, J. O.: Mechanisms of bone loss associated with total hip replacement. Orthop. Clin. North America, 24: 583-590, 1993.

Kelsey, J. L.; Thompson, W. D.; and Evans, A. S.: Methods in Observational Epidemiology, pp. 288-293. New York, Oxford University Press, 1986.

Kilgus, D. J.; Shimaoka, E. E.; Tipton, J. S.; and Eberle, R. W.: Dual-energy x-ray absorptiometry measurement of bone mineral density around porous-coated cementless femoral implants. J. Bone and Joint Surg., 75-B(2): 279-287, 1993.

Kronick, J. L.; Barba, M. L.; and Paproski, W. G.: Extensively coated femoral components in young patients. Clin. Orthop., 344: 263-274, 1997.

Lachman, E.: Osteoporosis: the potentialities and limitations of its roentgenologic diagnosis. Am. J. Roentgenol., 74: 712-715, 1955.

Lenke, L. G.; Betz, R. R.; Bridwell, K. H.; Clements, D. H.; Harms, J.; Lowe, T. G.; and Shufflebarger, H. L.: Intraobserver and interobserver reliability of the classification of thoracic adolescent idiopathic scoliosis. J. Bone and Joint Surg., 80-A: 1097-1106, Aug. 1998.

McAuley, J. P.; Culpepper, W. J.; and Engh, C. A.: Total hip arthroplasty. Concerns with extensively porous coated femoral components. Clin. Orthop., 355: 182-188, 1998.

McCarthy, C. K.; Steinberg, G. G.; Agren, M.; Leahey, D.; Wyman, E.; and Baran, D. T.: Quantifying bone loss from the proximal femur after total hip arthroplasty. J. Bone and Joint Surg., 73-B(5): 774-778, 1991.

McGovern, T. F.; Engh, C. A.; Zettl-Schaffer, K.; and Hooten, J. P.: Cortical bone density of the proximal femur following cementless total hip arthroplasty. Clin. Orthop., 306: 145-154, 1994.

Maloney, W. J.; Sychterz, C.; Bragdon, C.; McGovern, T.; Jasty, M.; Engh, C. A.; and Harris, W. H.: Skeletal response to well fixed femoral components inserted with and without cement. Clin. Orthop., 333: 15-26, 1996.

Matsui, M.; Nakata, K.; Masuhara, K.; Ohzono, K.; Sugano, N.; and Ochi, T.: The Metal-Cancellous Cementless Lübeck total hip arthroplasty. Five-to-nine-year results. J. Bone and Joint Surg., 80-B(3): 404-410, 1998.

Moreland, J. R., and Bernstein, M. L.: Femoral revision hip arthroplasty with uncemented, porous-coated stems. Clin. Orthop., 319: 141-150, 1995.

Mulliken, B. D.; Bourne, R. B.; Rorabeck, C. H.; and Nayak, N.: A tapered titanium femoral stem inserted without cement in a total hip arthroplasty. Radiographic evaluation and stability. J. Bone and Joint Surg., 78-A: 1214-1225, Aug. 1996.

O'Keeffe, D.: Metabolic bone disease. Morphometry. Radiol. Clin. North America, 29: 165-174, 1991.

Piston, R. W.; Engh, C. A.; De Carvalho, P. I.; and Suthers, K.: Osteonecrosis of the femoral head treated with total hip arthroplasty without cement. J. Bone and Joint Surg., 76-A: 202-214, Feb. 1994.

Rosner, B.: Fundamentals of Biostatistics. Ed. 4, pp. 423-426. Belmont, California, Duxbury Press, 1995.

Sakalkale, D. P.; Eng, K.; Hozack, W. J.; and Rothman, R. H.: Minimum 10-year results of a tapered cementless hip replacement. Clin. Orthop., 362: 138-144, 1999.

Smith, S. W.; Meyer, R. A.; Connor, P. M.; Smith, S. E.; and Hanley, E. N., Jr.: Interobserver reliability and intraobserver reproducibility of the modified Ficat classification system of osteonecrosis of the femoral head. J. Bone and Joint Surg., 78-A: 1702-1706, Nov. 1996.

Svanholm, H.; Starklint, H.; Gundersen, H. J.; Fabricius, J.; Barlebo, H.; and Olsen, S.: Reproducibility of histomorphologic diagnoses with special reference to the kappa statistic. APMIS, 97: 689-698, 1989.

Wan, Z.; Dorr, L. D.; Woodsome, T.; Ranawat, A.; and Song, M.: Effect of stem stiffness and bone stiffness on bone remodeling in cemented total hip replacement. J. Arthroplasty, 14: 149-158, 1999.

Ward, W. T.; Vogt, M.; Grudziak, J. S.; Tumer, Y.; Cook, P. C.; and Fitch, R. D.: Severin classification system for evaluation of the results of operative treatment of congenital dislocation of the hip. A study of intraobserver and interobserver reliability. J. Bone and Joint Surg., 79-A: 656-663, May 1997.

West, J. D.; Mayor, M. B.; and Collier, J. P.: Potential errors inherent in quantitative densitometric analysis of orthopaedic radiographs. J. Bone and Joint Surg., 69-A: 58-64, Jan. 1987.

Woolson, S. T., and Maloney, W. J.: Cementless total hip arthroplasty using a porous-coated prosthesis for bone ingrowth fixation. 3-year follow-up. J. Arthroplasty, 7 (Supplement): 381-388, 1992.

Topics

View Large | Download Slide(.ppt)
Add to My JBJS

Fig. 2:: Illustration showing the mean bone loss from fourteen pairs of femora, as determined with dual-energy x-ray absorptiometry in each of the sixteen zones examined.

View Large | Download Slide(.ppt)
Add to My JBJS

View Larger

Table I: Grouped Dual-Energy X-Ray Absorptiometry Results

Group	Mean Bone Loss (percent)	Agreement (percent)			Kappa Coefficient	P Value
Bone Loss Not Present	Bone Loss Present	Total
1 (n = 32)	-14.9	68.8	3.1	71.9	0.32	0.12
2 (n = 32)	3.3	25.0	37.5	62.5	0.49	<0.01
3 (n = 32)	17.3	25.0	31.3	56.3	0.41	<0.01
4 (n = 32)	31.3	6.3	53.1	59.4	0.30	0.05
5 (n = 32)	50.1	0.0	75.0	75.0	0.11	0.71
6 (n = 32)	69.9	6.3	90.6	96.9	0.85	0.02
7 (n = 32)	96.3	6.3	87.5	93.8	0.73	0.02

Add to My JBJS

Table I: Grouped Dual-Energy X-Ray Absorptiometry Results

Group	Mean Bone Loss (percent)	Agreement (percent)			Kappa Coefficient	P Value
Bone Loss Not Present	Bone Loss Present	Total
1 (n = 32)	-14.9	68.8	3.1	71.9	0.32	0.12
2 (n = 32)	3.3	25.0	37.5	62.5	0.49	<0.01
3 (n = 32)	17.3	25.0	31.3	56.3	0.41	<0.01
4 (n = 32)	31.3	6.3	53.1	59.4	0.30	0.05
5 (n = 32)	50.1	0.0	75.0	75.0	0.11	0.71
6 (n = 32)	69.9	6.3	90.6	96.9	0.85	0.02
7 (n = 32)	96.3	6.3	87.5	93.8	0.73	0.02