Abstract
Background: International registries with large, heterogeneous
patient populations provide excellent research opportunities for studying
factors that influence treatment outcomes after total hip arthroplasty. In the
present study, we used a European multinational database to investigate
whether there is an association between three functional variables
(preoperative pain, mobility, and motion) and functional outcome.
Methods: We performed a retrospective cohort study on preoperative
and follow-up clinical data that were prospectively entered into the
International Documentation and Evaluation System European hip registry
between 1967 and 2002. The inclusion criteria for this study were an age of
more than twenty years, an underlying diagnosis of osteoarthritis, and a
Charnley class-A functional designation at the time of surgery. A total of
12,925 patients (13,766 total hip arthroplasties) who met these criteria were
entered into the analysis. Three functional variables (pain, mobility, and
motion) that were assessed preoperatively were evaluated postoperatively at
various follow-up examinations for a maximum of ten years.
Results: Six thousand four hundred and one patients could walk
longer than ten minutes preoperatively; of these, 57.1% had a walking capacity
of more than sixty minutes at the time of the most recent follow-up. In
comparison, 6896 patients had a preoperative walking capacity of less than ten
minutes and only 38.9% of these patients could walk more than sixty minutes at
the time of the most recent follow-up. The difference was significant (p <
0.01). Similarly, 10,375 patients had a preoperative hip flexion range of
>70°; of these, 74.7% had a flexion range of >90° at the time of
the most recent follow-up. In comparison, 2793 patients had a preoperative hip
flexion range of <70° and only 62.6% of these patients had a flexion
range of >90° at the time of the most recent follow-up. The difference
was also significant (p < 0.01). Lasting, complete, or almost complete pain
relief was achieved by >80% of the patients following total hip
arthroplasty regardless of their preoperative categorization of pain.
Conclusions: Patients with poor preoperative walking capacity and
hip flexion are less likely to achieve an optimal outcome with regard to
walking and motion. In contrast, there is no correlation between the
preoperative pain level and pain alleviation, which is generally good and
long-lasting after total hip arthroplasty.
Level of Evidence: Prognostic Level II. See Instructions
to Authors for a complete description of levels of evidence.
The primary indications for total hip arthroplasty are pain and loss of
function. Although attempts have been made to develop explicit indications
criteria for total hip
arthroplasty1, the
extent to which preoperative pain and function influence the ultimate outcome
remains unknown2.
Using the Rand scale, Braeken et al. found an association between the
preoperative pain status and the postoperative pain level but could not
confirm these findings with use of the Western Ontario and McMaster
Universities Osteoarthritis Index (WOMAC) outcomes
instrument3.
MacWilliam et al. studied the effects of variables such as race, education,
comorbidities, and preoperative functional status on treatment outcome but did
not quantify the
results4. Fortin et
al. reported lower postoperative scores in the SF-36 physical domains and the
WOMAC when the preoperative scores were also
low5.
The goal of this multicentric retrospective European cohort study was to
assess pain, walking capacity, and hip flexion preoperatively and to determine
the association of each of these factors with ultimate function.
The community of hospitals contributing data to the
IDES6,7
(International Documentation and Evaluation System) hip registry consists of
academic and nonacademic centers, both private and public. A total of
sixty-five hospitals from eight European countries (Switzerland, Austria,
Germany, France, Belgium, Italy, The Netherlands, and Spain) have
prospectively collected standardized information in the first endeavor to
implement a pan-European, post-market surveillance system for total hip
arthroplasty. The majority of participants are located in Switzerland,
Germany, France, and Italy. The registry began in 1965 in Bern, Switzerland,
and data collection is still ongoing. Data for the study were recorded on
paper forms (see Appendix), sent to a central office in Bern, scanned with
optical mark readers, and stored in a relational database. Data entry
validation procedures ensured that all required data were included. The
variables recorded and terminology used on the IDES forms are based on
Clinical and Radiographic Terminology
(CART)8, a standard
system for reporting the results of total hip arthroplasty that represents a
consensus from the American Academy of Orthopaedic Surgeons (AAOS), the
International Society of Orthopaedic Surgery and Traumatology (SICOT), and the
Hip Society.
The current study did not require institutional review board approval at
our center as it utilized existing anonymous observational data. We queried
the registry database in the year 2005 to recruit all records for patients who
underwent primary total hip arthroplasty, had one or more complete follow-up
examinations, and met three inclusion criteria: an age of more than twenty
years, a diagnosis of osteoarthritis, and ipsilateral involvement of the hip
at the time of the primary total hip arthroplasty. No patient who met these
three criteria was excluded. Patients with unilateral hip disease who had
medical comorbidities sufficient to compromise walking capacity were assigned
to Charnley class C and therefore were excluded from the study.
Sample Characteristics
At the time that we initiated this study, there were 42,739 patients
(49,865 primary total hip arthroplasties) in the registry. Of these, 12,925
patients (13,766 total hip arthroplasties) met the inclusion criteria. The
complete sampling procedure is shown in the Appendix. The first patient in the
present study underwent total hip arthroplasty in 1967, and the most recent
follow-up information was entered into the database in 2002. The study group
included 6467 female patients and 6458 male patients with mean ages at the
time of surgery of 68.6 years (range, 24.3 to 94.8 years) and 66.3 years
(range, 22.8 to 94.7 years), respectively. The shortest duration of
postoperative follow-up was twenty-nine days, and the longest was twenty-seven
years. Because of the low numbers of follow-up examinations after ten years,
the analysis was restricted to a maximum follow-up period of ten years. When
multiple follow-up visits were recorded in the same year, the one closest to
the middle of the year was chosen and the others were discarded, leaving a
total of 27,712 follow-up records that were evaluated. The number of annual
follow-up assessments per year for the 13,766 primary total hip arthroplasties
is shown in the electronic Appendix. The average number of follow-up
assessments per patient was 2.1 (range, one to nine), and the mean duration of
follow-up was 4.3 years (range, twenty-nine days to ten years).
Table I shows the IDES
variables8 that were
recorded for the current study.
For reporting the results of the present study, we then combined certain
categories of pain and function. Pain was classified as none/mild, moderate,
or severe/intolerable; walking capacity was classified as more than sixty
minutes, thirty-one to sixty minutes, ten to thirty minutes, or less than ten
minutes/not possible; and the range of hip flexion was classified as
>90°, 71° to 90°, 30° to 70°, or <30°/stiff.
We also defined an excellent outcome as no or mild hip pain, a walking
capacity of greater than sixty minutes, and a range of hip flexion of
>90°.
We used a modification of "mean age related ability"
(MARA)9 curves to
show the relationships between preoperative pain and function and
postoperative functional outcome.
Statistical Analysis
All follow-up examinations were grouped on the basis of follow-up interval.
Generalized estimation equations were used to test the influence of patient
characteristics (gender, age at the time of the operation, preoperative
status) and time after surgery on outcome. For each outcome variable and each
follow-up year, a Cochran-Mantel-Haenszel test was used to compare the
preoperative patient groups. The Cochran-Mantel-Haenszel statistics for the
ten follow-up years were stratified for gender and age (less than sixty years,
sixty to seventy years, and more than seventy years).
Bonferroni-Holm adjustments for each outcome variable were set to account
for multiple testing over the ten follow-up years. Ninety-five percent
confidence intervals were used to graphically display the variability within
the groups. Linear regression was used to assess trends between categories of
preoperative walking and the average age at the time of the operation. The
effect size of age on preoperative walking was estimated with use of Cohen's
d. Age differences between groups with preoperative walking capacities of
greater than sixty minutes and less than ten minutes and between groups with
preoperative hip flexion ranges of >90° and <30° were calculated
with the Student t test. Correlations between walking capacity and range of
hip flexion were assessed with use of Kendall's tau b. The level of
significance was set at 0.05 throughout the study. All statistical analyses
were conducted with use of SAS 9.1 (SAS Institute, Cary, North Carolina).
Preoperatively, more than half of the patients had severe or intolerable
pain (Table II), could walk
only less than ten minutes without walking aids
(Table III), and had an active
range of hip flexion of 71° to 90°
(Table IV).
Hip Pain
Pain relief was almost immediate. Figure
1 shows the relationship between preoperative pain and
postoperative pain over time. The proportion of patients with complete or
almost complete alleviation of pain remained >80% for all three
preoperative pain groups at ten years postoperatively. During the complete
follow-up period, there was no significant difference in the proportion of
pain-free patients in any of the preoperative pain categories.
Walking Capacity
Figure 2 shows large and
significant (p < 0.01) differences among the four preoperative groups in
terms of the proportion of patients with a walking capacity of greater than
sixty minutes postoperatively. Six thousand four hundred and one patients
could walk longer than ten minutes preoperatively; of these, 57.1% had a
walking capacity of more than sixty minutes at the time of the most recent
follow-up. In comparison, 6896 patients had a preoperative walking capacity of
less than ten minutes and only 38.9% of these patients could walk more than
sixty minutes at the time of the most recent follow-up. The difference was
significant (p < 0.01). All groups demonstrated improved walking capacity
until three to four years postoperatively and then had a slow but constant
decline thereafter. In the group of patients who had a preoperative walking
capacity of greater than sixty minutes, >75% retained that walking capacity
at one year after surgery. Ninety percent could walk for more than sixty
minutes in the third postoperative year, and then there was a gradual decline
in walking capacity over the next seven years. Overall, the patterns of
postoperative gain and loss of walking capacity were similar for all four
groups. The average proportion of patients with postoperative walking capacity
of greater than sixty minutes was about 65% in the third year after surgery.
Patients with the worst preoperative walking capacity also had the worst
postoperative recovery of walking capacity. The average age difference between
patients with excellent preoperative walking capacity and those with the
poorest preoperative walking capacity was five years (p < 0.01). Linear
regression revealed that patients were an average 1.6 years older per category
of reduced preoperative walking capacity (p < 0.01). Cohen's d as a measure
of effect size was 0.4, indicating that age had a moderate effect on
deterioration of preoperative walking capacity. In order to assess the
influence of comorbidities on postoperative walking capacity, we analyzed the
percentage of patients who had been categorized as Charnley class A
preoperatively but came to be categorized as Charnley class C during the
follow-up period. The analysis was conducted for each category of preoperative
walking capacity for the second, fourth, sixth, eighth, and tenth follow-up
years. Table V reveals that
there was a consistent increase in the percentage of Charnley class-C patients
with each decrease in category of preoperative walking capacity at each of
these follow-up years.
Range of Hip Flexion
Figure 3 shows that, with
regard to hip flexion, the best outcome in all preoperative groups was
observed in the second, third, and fourth postoperative years. In the third
postoperative year, an average of 85% of the patients had an excellent range
of hip flexion that was followed by a slow but constant decline over the next
seven years. There were large and significant (p < 0.01) differences among
the four preoperative groups. Similarly, 10,375 patients had a preoperative
hip flexion range of >70°; of these, 74.7% had a flexion range of
>90° at the time of the most recent follow-up. In comparison, 2793
patients had a preoperative flexion range of <70° and only 62.6% of
these patients had a flexion of >90° at the time of the most recent
follow-up. The difference was also significant (p < 0.01). Of interest, the
group with an excellent range of hip flexion preoperatively sustained a slight
loss of flexion range after surgery. The pattern of postoperative gain and
loss of hip flexion range was similar for all four groups. Patients with
preoperative flexion of =70° had the worst postoperative recovery of
motion. In contrast to the findings regarding walking capacity, patients with
an excellent preoperative hip range of motion were an average three years
older than those with the poorest preoperative motion (p < 0.01).
The correlation between hip flexion range and walking capacity was only
weak (Kendall's tau = 0.099; 95% confidence interval, 0.087 to 0.11).
In the present study, a large number of European patients who had undergone
total hip arthroplasty were analyzed in terms of three variables: pain,
walking capacity, and range of hip
flexion10.
We found enduring, excellent relief of pain regardless of the preoperative
pain status. Approximately 95% of the study sample had no hip pain or only
mild hip pain in the first postoperative year, and the proportion of pain-free
patients declined only slightly over the ten-year study period. The
consistency of freedom from pain after total hip arthroplasty is in agreement
with the findings of many previous studies and is one of the main reasons for
the success of this
intervention11-13.
In contrast, postoperative walking capacity was significantly influenced by
the preoperative walking capacity. The greater the preoperative walking
capacity, the higher the proportion of patients with excellent postoperative
walking capacity. In addition, the preoperative range of flexion had a
dramatic influence on postoperative flexion. The groups of patients with a
preoperative range of hip flexion of =70° had the worst postoperative
flexion.
Our findings are in agreement with those of other authors who found
associations between the preoperative and postoperative functional
status2,5.
In addition, our findings are in agreement with those of MacWilliam et
al.4, who found that
patients with excellent preoperative pain and function scores experience an
initial postoperative decline. In contrast to Fortin et
al.5, we found a
constant relative improvement in outcome during the first three to four
postoperative years, even in patients with poor preoperative values. We cannot
fully explain why patients with poor preoperative function do not achieve
better functional status after total hip arthroplasty. While this group of
patients achieved substantially more relative improvement after surgery when
compared with patients with good preoperative
function4, their
absolute outcome remained suboptimal. One obvious factor compromising their
recovery is increasing age. In a previous study, we showed that the proportion
of patients with postoperative walking capacities of greater than sixty
minutes was greatest among those who were twenty to fifty years old, lower
among those who were fifty to seventy years old, and lowest among those who
were more than seventy years
old14. In the
current study, there also was a significant increase in average age for each
decrease in category of preoperative walking capacity. Cohen's d, a rather
conservative measure of effect size, revealed a moderate influence of age on
preoperative walking. In accordance with our findings, Papadakis et al.
studied healthy older men and concluded that age is the most important
variable in predicting functional
decline15.
Patients with a preoperative range of hip flexion of =70° were an
average of two years younger than those with a preoperative range of flexion
of >90°. Considering the weak associations between mobility and motion
and the difference in ages between the groups with the best and the worst
preoperative ranges of motion, age cannot be seen as an important influential
factor on postoperative motion. As pain alleviation was similar for all
patients and associations between mobility and motion were only weak, the
age-related decline in the functional status of older patients is a likely
explanation for their compromised postoperative performance. This conclusion
is substantiated by the frequency of patients who deteriorated from a Charnley
functional status A preoperatively to functional class-C during the follow-up
period. Table V reveals that
there was an increase in the percentage of Charnley class-C patients with each
decrease in category of preoperative walking capacity, and this increase was
very stable for the different follow-up periods. Hence, the older age of
patients at the time of surgery leads to a compromised recovery of mobility
because of rapidly emerging comorbid conditions. This is reflected in a lower
number of patients with excellent walking capacities in the respective
classes.
The present study had some limitations. It investigated the treatment
outcome following total hip arthroplasties performed with different component
designs and fixation modes and included patients from multiple centers and
surgeons with different levels of experience. All of these factors could have
influenced the study findings. In addition, although the study followed
patients through the tenth postoperative year, patients usually did not have a
complete record of ten documented follow-up examinations and the analysis did
not account for clustering of data by center of treatment. Thus, the study
results may be positively biased by the withdrawal of patients who had an
undesired outcome from follow-up routines at the center of the primary
intervention. On the other hand, we controlled the Cochran-Mantel-Haenszel
statistics in this cohort study for the covariables of gender, age, and
follow-up year. In addition, the statistical analyses accounted for multiple
tests over time, and, despite losses to follow-up in each year, the absolute
number of observations was sufficiently large to provide meaningful estimates
and confidence intervals.
In spite of these limitations, our study benefits from major advantages,
which include the consistent methodology and terminology of the
data-collection instruments and the fact that results span nearly forty years
of data from a multitude of centers—both academic and
nonacademic—in eight European countries. This enhances the possibility
of generalizing our findings.
These strengths and weaknesses also reflect the pros and cons of analyzing
large-scale, international registries. Scandinavian registries have
demonstrated that small data sets that are routinely collected for primary and
revision procedures are sufficient for calculating valid and reliable implant
survival rates if they measure a high percentage of interventions. However,
such data sets are less suitable for analyses of clinical interrelationships
because they lack detailed documentation of patient characteristics and
treatment outcomes, especially at times when the implants are well
functioning, i.e., at routine follow-up examinations.
In conclusion, our results indicate that poor preoperative mobility and
function are predictors of less postoperative recovery of function in patients
who are categorized as Charnley class A, whereas pain is insensitive to
preoperative level and is generally alleviated with good and long-lasting
effect.
Figures showing the International Documentation and Evaluation System
primary and follow-up forms and the selection process of patients from the
registry, and a table showing the number of annual follow-up assessments, are
available with the electronic version of this article, on our web site at
(go to
the article citation and click on "Supplementary Material") and on
our quarterly CD-ROM (call our subscription department, at 781-449-9780, to
order the CD-ROM). ?
Quintana JM, Arostegui I, Azkarate J,
Goenaga JI, Elexpe X, Letona J, Arcelay A. Evaluation of explicit criteria for
total hip joint replacement. J Clin Epidemiol.
2000;53:
1200-8.531200
2000
[PubMed][CrossRef]
Holtzman J, Saleh K, Kane R. Effect of
baseline functional status and pain on outcomes of total hip arthroplasty.
J Bone Joint Surg Am.
2002;84:
1942-8.841942
2002
[PubMed]
Braeken AM, Lochhaas-Gerlach JA, Gollish
JD, Myles JD, Mackenzie TA. Determinants of 6-12 month postoperative
functional status and pain after elective total hip replacement. Int J
Qual Health Care. 1997;9:
413-8.9413
1997
MacWilliam CH, Yood MU, Verner JJ,
McCarthy BD, Ward RE. Patient-related risk factors that predict poor outcome
after total hip replacement. Health Serv Res.
1996;31:
623-38.31623
1996
[PubMed]
Fortin PR, Penrod JR, Clarke AE,
St-Pierre Y, Joseph L, Belisle P, Liang MH, Ferland D, Phillips CB, Mahomed N,
Tanzer M, Sledge C, Fossel AH, Katz JN. Timing of total joint replacement
affects clinical outcomes among patients with osteoarthritis of the hip or
knee. Arthritis Rheum.
2002;46:
3327-30.463327
2002
[PubMed][CrossRef]
Paterson D. The International
Documentation and Evaluation System (IDES). Orthopedics.
1993;16:
11-4.1611
1993
[PubMed]
Röder C, Eggli S, El-Kerdi A,
Müller U, Ambrose T, Roosli E, Busato A, Aebi M. The International
Documentation and Evaluation System (IDES)—10-years experience.
Int Orthop. 2003;27:
259-61.27259
2003
[PubMed][CrossRef]
Johnston RC, Fitzgerald RH Jr, Harris
WH, Poss R, Müller ME, Sledge CB. Clinical and radiographic evaluation of
total hip replacement. J Bone Joint Surg Am.
1990;72: 161-8.
Erratum in: J Bone Joint Surg Am. 1991;73:952.72161
1990
[PubMed]
Dubs L. Der Patient als Experte.
[Patients as experts: determining benefit by using assessments of ability
(ICIDH)]. Forsch Komplementärmed.
1998;5 Suppl S1:
90-5. German.590
1998
[PubMed][CrossRef]
Bryant MJ, Kernohan WG, Nixon JR, Mollan
RA. A statistical analysis of hip scores. J Bone Joint Surg Br.
1993;75:
705-9.75705
1993
[PubMed]
Jones CA, Voaklander DC, Johnston DW,
Suarez-Almazor ME. Health related quality of life outcomes after total hip and
knee arthroplasties in a community based population. J
Rheumatol. 2000;27:
1745-52.271745
2000
March LM, Cross MJ, Lapsley H, Brnabic
AJM, Tribe KL, Bachmeier CJ, Courtenay BG, Brooks PM. Outcomes after hip or
knee replacement surgery for osteoarthritis. A prospective cohort study
comparing patients' quality of life before and after surgery with age-related
population norms. Med J Aust.
1999;171:
235-8.171235
1999
[PubMed]
Wiklund I, Romanus B. A comparison of
quality of life before and after arthroplasty in patients who had arthrosis of
the hip joint. J Bone Joint Surg Am.
1991;73:
765-9.73765
1991
[PubMed]
Röder C, Parvizi J, Eggli S, Berry
DJ, Müller ME, Busato A. Demographic factors affecting long-term outcome
of total hip arthroplasty. Clin Orthop Relat Res.
2003;417:
62-73.41762
2003
[PubMed]
Papadakis MA, Grady D, Tierney MJ, Black
D, Wells L, Grunfeld C. Insulin-like growth factor 1 and functional status in
healthy older men. J Am Geriatr Soc.
1995;43:
1350-5.431350
1995
[PubMed]