Elective total hip replacement: incidence, emergency readmission rate, and postoperative mortality

OBJECTIVES

To report the incidence of elective total hip replacement and postoperative mortality, emergency readmission rates, and the demographic factors associated with these rates in a large defined population.

DESIGN

Analysis of linked, routine abstracts of hospital inpatient records and death certificates.

SETTING

10 hospitals in six districts in Oxford Regional Health Authority covered by the Oxford record linkage study.

SUBJECTS

Records for 11,607 total hip replacements performed electively in 1976-85.

MAIN OUTCOME MEASURES

Incidence of operation, postoperative mortality, relative mortality ratios, and incidence of emergency readmission.

RESULTS

NHS operation rates increased over time from 43 to 58 operations/100,000 population. Variation in operation rates between districts reduced over time. Operation rates were on average 25% higher in women than men. There were 93 deaths (11/1000 operations) within 90 days of the operation and 208 emergency readmissions (28/1000 operations) within 28 days of discharge. Postoperative mortality and emergency readmission rates increased with age. No significant trend with time was found. Mortality in the 90 days after the operation was 2.5-fold higher (1.9 to 3.0) than in the rest of the first postoperative year. This represented an estimated excess of 6.5 (4.2 to 8.8) early postoperative deaths/1000 operations. Most deaths were ascribed to cardiovascular events. Thromboembolic disease was the commonest reason for emergency readmission.

CONCLUSIONS

The pronounced increase in operations in districts with initially low rates suggests a trend towards greater equity in the local provision of NHS hip arthroplasty. The early postoperative clusters of deaths attributed to cardiovascular disease and of readmissions for thromboembolic disease suggest that there is scope for investigating ways of reducing the incidence of major adverse postoperative events.

Registration of arthroplasties in Finland. A nationwide prospective project

Data on hip and knee arthroplasties have been compiled on a nationwide basis in Finland since 1980. Forty-five major departments contribute to the study providing data on the type of operation, the implant used, the diagnosis, and the 1-year clinical results. In the case of revision, new data are sent to the register, enabling survivorship analysis. Between 1980 and 1988, 25,966 operations were reported. Fifty-six percent had been made for primary osteoarthrosis, 22 percent for rheumatoid arthritis, 6.3 percent for secondary arthrosis, and 0.5 percent for CDH. In 1988, the total number of arthroplasties was 4,628: about two thirds hip and one third knee replacements. The annual incidence of primary total hip arthroplasties in 1988 was 58 per 100,000 inhabitants and that for the knees 25 per 100,000 inhabitants. More than 40 percent of the patients were under 65 years of age. In the whole series, primary thromboembolic complications occurred in 1.4 percent, luxations in 1.4 percent, infection in 0.9 percent, and evacuated hematoma in 0.6 percent. The annual frequency of re-arthroplasty increased between 1980 and 1988 from 9.8 to 13.6 percent, indicating an increasing orthopedic work load in the future.

Cup containment and orientation in cemented total hip arthroplasties

We reviewed the radiographs of 864 Charnley and STH (Zimmer) cemented total hip arthroplasties with a mean follow-up of seven years (maximum 16 years). Survivorship analysis was used to assess the correlation between radiographic performance and the bony containment or the coronal orientation of the acetabular cup. The cup orientation and containment were interrelated; all vertically oriented cups were completely contained, whereas 25% of more horizontal cups were only partially contained. Completely contained cups had significantly lower incidences of complete cement-bone radiolucency (p = 0.02) and of wear (p = 0.09). Vertically oriented cups had a lower incidence of continuous radiolucency than neutrally oriented cups, but this was not statistically significant (p = 0.25). Our results confirm the importance of complete bony containment, and also indicate that it is better to accept vertical orientation and obtain full bony coverage than to have a more horizontal orientation with partial containment.

[Use of economic incentives to promote hip replacement surgery at the regional hospital in Tromsö]

Since 1 December 1988, the Norwegian Health Department has refunded NOK 12,000 for each patient operated with a total hip replacement. The intention has been to encourage the hospitals to increase their capacity to operate such patients. The Regional Hospital of Tromsø has paid the hospital staff extra to operate patients on Saturdays. In this way 58 extra patients received a total hip replacement in 1989. The cost for each such extra patient was NOK 24,869. In comparison, the regular cost for such patients is NOK 35,125. In our opinion, the hospital has found a simple and economical way to increase its capacity for operating patients with total hip replacement.

Survival of cemented total hip replacements

A Kaplan-Meier survival analysis of 240 Charnley, 149 Mueller (Depuy, Warsaw, IN), 576 Trapezoidal 28 (Zimmer, Warsaw, IN), and 100 Miami Orthopaedic Surgical Clinics (Biomet, Warsaw, IN) cemented total hip replacements with all polyethylene acetabular components, performed by the senior author (M.A.R.) should be the basis that all noncemented total hip replacements must exceed to be a reasonable alternative. Failure, ie, revision and a loose prosthesis, shows that 80%, 61%, and 63% of the Charnley, Mueller, and Trapezoidal 28, respectively, survive longer than 15 years. The Miami Orthopaedic Surgical Clinics prosthesis, which uses current cement technology, has been observed for only 8 years. Despite the fact that there are no differences noted between all four prostheses at 8 years, more time is needed to judge the usefulness of current cement technology.

Long-term survival of the Charnley low-friction total hip arthroplasty

A survivorship analysis of 1,151 Charnley low-friction total hip arthroplasties performed as a primary procedure showed the following 15-year failure probabilities: hip revision 13%, cup revision 8%, femoral revision 12%, radiographic femoral loosening 13%, and radiographic cup loosening 49%. Pain scores diminished gradually with time so that at 15 years, the probability of a Charnley pain score less than 5 was 70%. The data suggest that the Charnley low-friction hip arthroplasty has a finite life span.

Blood conservation in major orthopedic surgery

During an eight-year period, intraoperative autotransfusion was performed in 1922 operations. There were 476 primary hip arthroplasties, 1017 revision hip arthroplasties, 339 spinal procedures, and 90 vascular cases. There were no known complications related to autotransfusion in any of the cases. With proper technique, red cell salvage is approximately 58% effective, and intraoperative blood loss can be reduced by that amount. The first four years' experience with the preoperative autogeneic blood program is as follows. Using a combined program of predonation and primary hip arthroplasty, 72% of the revision hip arthroplasties and 81% of the spinal instrumentations required only autologous blood. A combined program of autologous donation and intraoperative autotransfusion is an excellent alternative to allogeneic blood replacement and a means of eliminating transfusion-related disease.

Anatomic Porous Replacement hip arthroplasty: first 100 consecutive cases

One hundred consecutive Anatomic Porous Replacement (Intermedics Orthopaedics, Austin, TX) hip replacements were followed for 4 years. Eighty percent of patients had bone ingrowth fixation, 14% stable fibrous, and 6% unstable fibrous (loose) with 4% revised. Only two hips changed fixation grade after 2 years. Bone remodeling showed cancellous hypertrophy of the cortex, usually along the lateral cortex, and 7% had stress shielding (atrophy) of the proximal cortices. Adaptive bone remodeling almost always occurred in type B and C bone. Bone remodeling was statistically related to bone type, prosthetic fill, stem-bone ratio, and collar loading. We concluded that proximal bone ingrowth fixation with proximal load transfer provides good clinical results and excellent bone remodeling. Also, collar loading improves bone response. To expand this fixation in a predictable fashion to all type B and some type C bone requires geometry changes, which have been accomplished in the Anatomic Porous Replacement II.

Survival of a cemented porous-coated prosthesis

Porous-surfaced femoral components were approved by the United States Food and Drug Administration for implantation with polymethyl methacrylate cement in total hip arthroplasty. The ability of porous coating to prevent loosening of prostheses in vivo has not been reported in the literature previously. This study compared the radiographic and clinical success of two structurally similar prosthetic designs, one of which incorporates a porous coat. Kaplan-Meier survival analysis showed no significant difference between the components (P greater than .05) when failure is defined as revision alone, radiolucency alone, or the combination of revision and radiolucency.

Stem fracture incidence in Trapezoidal-28 stainless steel hip arthroplasty

From 1970 to March 1978, 716 forged stainless steel Trapezoidal-28 (T-28) total hip arthroplasties were performed at the authors' institution with a follow-up period from one to 15 years (mean, 64 months). There were 13 (1.8%) stem fractures and one (0.1%) neck fracture. Average time to fracture was 95 months. All fractures occurred in stems implanted before 1976, using early femoral acrylic stem fixation techniques. Stem fracture was related to time of surgery, stem size, patient height, patient weight (for males), and patient age. Survivorship analysis predicted an overall ten-year fracture rate of 4%, and a ten-year rate of 5% for T-28 stems implanted before 1976. There have been no reported fractures in the 204 T-28 stainless steel prostheses implanted since 1976. The mean patient follow-up period for the T-28 stems implanted after 1975 was 52 months, and only 16 patients have been followed for nine years or more. Thus, a longer patient follow-up period is needed in the post-1975 group in order to determine the long-term fracture rate in this group. However, the major difference between this group and the early group is the use of an improved cementation technique, as well as the use of larger prostheses when possible. There was no material or manufacturing change until the introduction of the TR-28 prostheses in 1978.

Survivorship analysis of 1,041 Charnley total hip arthroplasties

Survivorship analysis of 1,041 cemented Charnley total hip arthroplasties performed as a primary procedure revealed a probability of component survival at 10 years of 92%; the probability of acetabular cup survival was 99% and of femoral component survival was 96%. Three-zone acetabular demarcation was present in 16% of cases, as was migration of the cup greater than 5 mm. However, the acetabular revision rate was 1.65%, confirming the long-term clinical durability of the 22-mm internal diameter cup. Radiographic evidence of definite femoral component loosening was present in 9.6% and high-grade femoral bone-cement demarcation was present in 3.5%. The isolated femoral revision rate was 1.8%. Based on detailed survivorship analysis, a high-risk group of patients was identified for component failure and for femoral component loosening (radiographic). These patients were male, young (less than 50 years), heavy (greater than 170 pounds), and active (not Charnley class C). Given these findings, it is difficult to justify the widespread use of noncemented total hip systems, except in identifiable high-risk patients.

Total hip replacements in the National Health Service: is need being met?

The extent to which the need for total hip replacement during the early 1980s was being met by the NHS was assessed by examining patterns of utilization during the period 1978 to 1985 in and by residents of North West Thames region. During this period the number of operations performed on residents in NHS hospitals increased by 38 per cent. Within the private sector the number of total hip replacements performed on residents increased from 220 in 1981 to 720 in 1986. The amount of interdistrict variation in the rate of surgery declined as orthopaedic services became more widely available. The proportion of patients waiting more than one year declined from 14 to 4 per cent, suggesting an increasing ability for health services to meet need. Overall these findings suggest that despite improvements in provision during the early 1980s, significant unmet need still existed in 1985.

Cementless total hip arthroplasty using the anatomic medullary locking stem. Results using a survivorship analysis

Three hundred forty-three primary hip arthroplasties performed prior to June 1984 were followed for two years, and an additional 204 were followed for five years. Anteroposterior and lateral roentgenograms taken at annual intervals were used for comparison. At two years, 78% of the hips showed roentgenographic signs of stem osseointegration. Seventeen percent were clinically stable but demonstrated reactive lines surrounding the porous surface, indicating lack of bone-ingrowth fixation. Four percent showed late implant migration. Using late migration as an end point, a survivorship analysis of the stem stability was performed. The cumulative survival rate for stable fixation was 94% at five years and 88% at eight years. Among the 143 undersized stems, 17 showed late migration and two failed by stem fracture at six and eight years, giving a cumulative survival rate of 77% at eight years. By contrast, there were no roentgenographic failures among 200 canal-filling stems. Three stems were revised: two for stem fracture and one for loosening with infection. Proximal stress shielding occurred only in bone-ingrowth stems and was slightly progressive. Unstable implants produced progressive intramedullary canal widening. The clinical results were superior when the roentgenograms showed signs of bone-ingrowth fixation. Functional recovery in such cases was equivalent to that for cemented arthroplasty.

Multicenter clinical trials and their value in assessing total joint arthroplasty

Multicenter clinical trials (MCT) have an important role in the assessment of total joint arthroplasty. The primary reason for such MCT is the need to have access to a large number of patients. In Sweden, failure after total hip arthroplasty has been recorded prospectively since 1979. Medical records from every reoperation are documented and computer analyzed. This MCT has shown that patient-related, surgical, and implant-related parameters are of importance to rates of failure. There were an increasing number of revisions after total hip arthroplasty in Sweden between 1979 and 1986. Male gender and young age increase the risk of revision. The primary diagnosis is very important for type of failure. The risk of deep infection is small but increases with the number of previous operations. Aseptic loosening has emerged as the main problem and is the cause for 74% of all revisions. Prosthetic design is of utmost importance for the rate of failure, and significant differences exist among different prosthesis types with respect to long-term survival. Finally, this MCT has shown that revisions for total hip arthroplasty in patients younger than 70 years of age eventually have failed in Sweden whenever the cemented technique has been used.

Survivorship analysis of failure pattern after revision total hip arthroplasty

Failure, defined as established indication for or performed re-revision of one or both components, was analyzed using survivorship methods in 306 revision total hip arthroplasties. The longevity of revision total hip arthroplasties was inferior to that of previously reported primary total hip arthroplasties. The overall survival curve was two-phased, with a late failure period associated with aseptic loosening of one or both components and an early failure period associated with causes of failure other than loosening. Separate survival curves for aseptic loosening of femoral and acetabular components showed late and almost simultaneous decline, but with a tendency toward a higher rate of failure for the femoral component. No differences in survival could be found between the Stanmore, Lubinus standard, and Lubinus long-stemmed femoral components. A short interval between the index operation and the revision and intraoperative and postoperative complications were risk factors for early failure. Young age was a risk factor for aseptic loosening of the femoral component. Intraoperative fracture of the femoral shaft was not a risk factor for secondary loosening. No difference in survival was found between primary cemented total arthroplasty and primary noncemented hemiarthroplasty.

Development of the Belfast orthopaedic information system

A team from Musgrave Park Hospital, Belfast, describe their success in reducing the hip waiting list and introducing an information system to manage the orthopaedic waiting list.

Considerations in the comparison of cemented and cementless total hip prostheses

Based on conservative models of the revision rates for cemented and uncemented hip arthroplasties, the required number of patients for a study to show significant uncemented superiority was determined using standard statistical techniques. A review of the literature was done to ascertain the revision rates for cemented total hip arthroplasties. Patients with osteoarthritis who were older than 50 years had an average revision rate of 1% per year, while those younger than 50 years had revision rates in excess of 2% per year. Specific conditions of the hip were also associated with increased failure rates, most notably the implantation of a revision arthroplasty for a previously failed total hip arthroplasty. Assuming that the uncemented hip had a long-term revision rate of 0.5% per year and an early lack of biologic fixation in 1% of patients, it was found that a 5-year study would have to include at least 2,800 patients (1,400 in each group) if the study population had a cemented failure rate of 1% per year and a total of 230 patients in those having a cemented failure rate of 2% per year. The patients needed for a 10-year evaluation would be 700 and less than 100, respectively. Altering the long- or short-term revision rate for uncemented total hip arthroplasty had relatively little effect on the study size. This analysis indicates that a comparison study would be best done in patients who have high failure rates with cemented total hip arthroplasties. Multicenter trials involving younger patients or those undergoing revision surgery are ideally suited for such an investigation.

Use and abuse of performance indicators

An audit was performed by this department after allegations by the regional health authority of low productivity. It was found that the health authority had underestimated the number of operations performed in 1983 by only 5%, but an inexact classification and grading of operations had led to errors in the performance indicators of 19.8% for the "weighted number of operations" and 34.5% for the "number of major operations per consultant." When the throughput of orthopaedic departments in districts was compared by the regional health authority it was found that such errors in performance indicators had been further compounded by the inconsistent use of population data and incorrect data on medical staffing. Medical practitioners and the health authorities are alerted to this amplification of inaccurate data. Other methods for assessing trauma and orthopaedic surgery are proposed, such as a simplification of the Office of Population Censuses and Surveys classification of surgical operations, grading operations based on time spent in the operating theatre, and provision of computer programs to code for diagnosis and operation when writing discharge summaries.