Causes of Death After Hip Arthroplasty in Primary Arthrosis

Transcription

1 The Journal of Arthroplasty Vol. 12 No Causes of Death After Hip Arthroplasty in Primary Arthrosis Tuomo Visuri, MD,* Pekka Pulkkinen, MSc, t Pekka Paavolainen, MD,$ Markku Koskenvuo, MD, and Kaj. B. Turula, MD* Abstract: The causes of death in 1,018 patients operated on for primary osteoarthrosis with cemented total hip arthroplasty (THA) were compared with those of age- and sex-matched orthopaedic control patients and those of the general population in Finland. The mean follow-up period was 12 years for the THA patients and I 1 years for the control patients. During the first 4 years after surgery, the mortality of the THA patients from circulatory diseases was significantly increased compared with that of the orthopaedic control patients; the number of deaths in patients with THA was 34 compared with 17 for orthopaedic control patients, the relative risk being 2.00 (95% confidence interval, ). During the 10-year period after the surgery, the relative risk of death of the THA patients compared with the orthopaedic control patients was 1.50 for death from circulatory diseases (95% confidence intervals, ), 0.42 for accidental death (95% confidence interval, ), 0.74 for death from cancer, and 0.77 for death from other causes. The observed numbers of deaths from circulatory diseases or by accidents for patients with THA during a postoperative time frame of 5 to 23 years did not differ from the numbers expected for an age- and sex-matched subgroup of the Finnish population. The number of deaths from cancer was less than expected (P =.046). Key words: total hip arthroplasty, cause of death, primary arthrosis. Total hip arthroplasty (THA) was first performed in Finland in I967 using the McKee-Farrar metalon-metal prosthesis (Howmet International, Brussels, Belgium). Since 1973, this type of prosthesis has been replaced by plastic-on-metal prostheses. At present, the annual number of TEAs in Finland is about 4,000 (82/100,000 inhabitants) [1]. Total hip arthroplasty significantly improves all health-related quality-of-life functions, including walking ability [2]. Moderate physical activity has been shown to reduce the number of cardiovascular deaths [3-5]. Could the improvement in the walking ability of patients with THA influence their risk of death from cardiovascular causes? Patients with THA are exposed to the corrosion and wear products of the hip implant, which can pose the risk of cancer [6]. This risk may be seen in the cancer mortality of patients with THA. The cardiovascular, cancer, and accident mortality of patients with THA are compared with that of matched orthopaedic control patients and that of the general population in Finland. Materials and Methods From the *Central Military Hospital and ~:Hospital Orton, Helsinki, Finland, and the Department of Public Health, -puniversities of HeIsinki and fturku, Finland. Reprint requests: Tuomo Visuri, MD, Central Military Hospital, RO. Box 50, 00301, Helsinki, Finland Churchill Livingstone Inc. The basic population comprised 1,863 patients over 50 years of age who had undergone a THA for primary coxarthrosis from 1967 to 1985 at the Orthopedic Hospital of the Invalid Foundation (Hospital Orton). The control patients were 397

2 398 The Journal of Arthroplasty Vol. 12 No. 4 June 1997 selected from 2,026 candidates over 50 years of age operated on from 1973 to 1985 at the same hospital for other common orthopaedic indications, using the National Hospital Discharge Registry. For each patient with THA, an age- (within 1 year exactly) and sex-matched control patient was selected. Only complete pairs were included in the study. The age differences ol the THA and control groups caused exclusion of 846 patients with THA from the study series; thus, the final population comprised 1,018 THA and orthopaedic patient pairs (Table 1). At the time of the operation, the mean age of 630 male patients was 61, and that of the remaining 1,406 female patients, 63. The McKee-Farrar whole chrome-cobalt prosthesis was used from 1967 to 1975, the Brunswik chrome-cobalt polyethylene (Waldemar Link, Hamburg, Germany) from 1972 to 1981, and the Lubinus chrome-cobalt polyethylene prosthesis (Waldemar Link) from 1977 to 1985 (Table 1 ). Of the matched control patients, 138 were operated on for knee menisci, 395 for lumbar spinal conditions (spinal stenosis and herniated intervertebral disk), and 485 for hallux valgus. All persons in Finland have been identified from January 1, 1967, with a unique 10-digit personal identification number. The number is composed of 6 digits (day of birth, month, year), supplemented by a so-called registration number (3 digits) and a check digit. With the help of these personal identification numbers, the causes and dates of death for patients with THA and control patients were obtained from the Central Statistical Office of Finland [71. Hospital discharges have been registered in Finland since The registry has been reliable since 1972 [8]. Thus, the control group was selected from 1973 onward. None of the control patients later received a joint prosthesis or any other metallic implant, as ensured by the personal identification number from the hospital discharge register of the Orthopedic Hospital of the Invalid Foundation or from the Finnish Endoprosthesis Registry [1]. Table 1. Analysis of Control and THA Patients by Age and Sex Patients and No. of Mean Percentage Prostheses Patients Age (y) Women Control patients 1, McKe e-farrar Brunswik Lubinus Total 2, No antithrombosis prophylaxis was given to any of the THA patients, but they were mobilized on the first postoperative day. The first surgical intervention (date and type of prosthesis) was regarded as the starting point, and the follow-up period ended on December 31, Cases of bilateral or of revision operations were not induded in the study material. The mean follow-up period for the THA patients was 12 years, and for the orthopaedic control patients, I 1 years. No patients were lost to follow-up evaluation. The expected ageand sex-adjusted number of deaths in the normal population was calculated from the statistics of the Central Statistical Office of Finland [9]. The causes of death were divided into four main groups: cardiovascular, cancer, acddent, and other. Patients with THA had 12,549 person-years and control patients 11,635 person-years during the whole follow-up period. The results were reported mainly for the first 10-year period after surgery. During that time, THA patients had 9,114 personyears (men, 2,685; women, 6,429), and the control patients had 8,925 person-years (men, 2,625; women, 6,276). The differences in the causes of mortality causes were evaluated using relative mortality risks between the THA and control patients and standardized mortality ratios in population comparisons. For both comparisons, 95% confidence intervals were calculated. P values in population comparisons were calculated from observed and expected numbers of deaths by a two-sided Poisson distribution. When statements of "no difference" were presented, the power of the statistical test was also given. Power is calculated as the probability of declaring a statistically significant difference (alpha <.05) assuming that the observed difference is true. No adjustments of alpha and beta probabilities for multiple looks at the data were made. Results During the follow-up period, a total of 567 patients died (316 THA patients and 251 controls). During the first 10-year period after surgery, the corresponding figures were 173 and 152. No differences existed for all causes of death between the patients with metal-on-metal THA and those with plastic-on-metal THA (power <. 1). Cardiovascular mortality of the THA patients was significantly higher (P =.009) compared with the control patients, especially during the first 4 follow-up years (P =.03) (Tables 2, 3.) The risk was equalized during the next 16 years (Fig. 1).

3 Causes of Death After Hip Arthroplasty Visuri et al. 399 Table 2. Risk of Mortality from Cardiovascular Diseases of THA and Nonimplant Orthopaedic Control Patients in 20 Years Follow-up Period (y) No. of Deaths Person-years THA Control THA Control Relative 95% Confidence Patients Patients Patients Patients Risk Interval I I I THA, total hip arthroplasty. Differences in the cardiovascular mortality were related mainly to coronary deaths. There were 129 cardiac deaths in the THA series compared with 82 (P =.002) in the control patients. The total numbers of cerebrovascular deaths were 50 in the THA group and 39 (P =.3) in the control group. Four patients in the THA group died because of thromboembolic complications, 1, 45, 49, and 172 months, respectively, after the operation. The control patients had no thromboembolic deaths (P =. 125). No significant difference (power =.27) existed between the number of cardiovascular deaths in patients with THAs and the expected number of deaths in the general population (Table 4). The group of control patients, however, had fewer cardiovascular deaths than expected (P =.0002) (Table 4). The cancer mortality of the THA patients was reduced (P =.046) compared with that of the normal population, but did not differ (power =.048) from that of the control patients (Tables 3, 4). The group of THA patients had fewer (but not significantly: P =.08, power =.48) accidental deaths than the group of control patients, for whom the risk of accidental death was also higher (P =.03) than that for the normal population (Tables 3, 4). No significant differences (power =.3) existed in the other causes of death between THA and control patients (Table 3). Discussion Reliable medical computerized systems and registers are a precondition to epidemiological studies in a larger population. Data from the Finnish Hospital Discharge Register are sufficient for the epidemiologic research of diseases. In an epidemiologic cohort of 57,000 persons in , the discharge register covered 78% of all hospital treatments; for strokes and myocardial infarctions, the figures were 82% and 85%, respectively [8]. The agreement between hospital discharge records and the written patient history for musculoskeletal diagnoses was 96%, and for cardiovascular diagnoses, 95%, in 1986 [10]. Registration of deaths in Finland is virtually complete, and death records were available for all THA and control patients. Although follow-up evaluation of the control patients was started later than that of the patients with metal-on-metal prostheses, both groups are otherwise comparable. Mortality and causes of Table 3. Number of Deaths, Mortality per 10,000, and Relative Risk of Death of THA and Nonimplant Orthopaedic Control Patients in the 10 Years After the Operation by Cause of Death No. of Deaths Mortality per 10,000 THA Control THA Control Relative 95% Confidence Cause of Death Patients Patients Patients Patients Risk Interval Cancer Circulatory diseases I Accidents Other causes Total THA, total hip arthroplasty.

4 400 The Journal of Arthroplasty Vol. 12 No. 4 June 1997 Table 4. Observed Number of Deaths, Expected Number of Deaths, and Standardized Mortality Ratios for THA and Nonimplant Orthopaedic Control Patients from the Finnish Population by Cause of Death Observed No. of Deaths Expected No. of Deaths* Standardized Mortality Ratio THA Control THA Control THA Control Causes of Death Patients Patients Patients Patients Patients 95 % CI Patients 95 % CI Cancer 63 t Circulatory diseases Accidents ll *Calculated according to the Finnish Statistics of Causes of Death.tP =.046 (compared with 81.4). ~:P =.002 (compared with 188). P =.04 (compared with 13.7). CI, confidence interval; THA, total hip arthroplasty. death did not differ between patients with metalon-metal (McKee-Farrar) prostheses and those with plastic-on-metal (Brunswik, Lubinus) prostheses. Ten-year survivorship of the present series was 85%, in McKee-Farrar patients, 82% in Brunswik patients, and 82% in Lubinus patients [11]. Generally, patients with THA are healthier than the normal population because ot the preoperative selection. Their life expectancy is better than that of the general population in short, medium, or a long follow-up periods, and their causes of death may differ from those of the general population [11-13]. Thus, the causes of death of the THA patients should be compared with those ot a cohort of similar general health. The ideal group for comparison would be unoperated coxarthrotic patients, whose general health corresponds to that of THA patients. In practice, it is difficult to create such a group lot long-term studies because osteoarthrosis is a chronic, deteriorating disease that will not ease without treatment. For the comparison, a group of patients was selected who had been operated on for other common orthopaedic disorders and whose names had been submitted to preoperative selection for surgery. Both groups studied had thus undergone preoperative selection. The orthopaedic control patients were even healthier than the THA patients; among them, there were significantly fewer cardiovascular deaths than among patients with THA or the normal population. There were more accidental deaths among them than among the THA patients, which indicates a more active lifestyle and better health. The late causes of death in THA patients have not been well studied. Holmberg reported the causes of death in 518 THA patients, operated on for primary o ca ~D > I I I I q Years after operation Lower 95% CL Relative Risk Upper 95% CL. I~ * -~ Fig. 1. Decreasing mortality risk of total hip arthroplasty patients. CL, confidence level (interval).

5 Causes of Death After Hip Arthroplasty Visuri et al. 401 arthrosis, after a 6-year postoperative period [8]. The causes were cardiovascular disease in 60%, cancer in 21%, and other causes in 19% of the patients [12]. The corresponding tigures for the present series are 65%, 19%, and 16% in 10 years. Reasons for the increased number of cardiovascular deaths in THA patients compared with the control patients are not known. In a Swedish series of 279 THA patients operated on for primary arthrosis, no increase in the morbidity from cardiovascular diseases was observed compared with the same number of age- and sex-matched controls subjects from the same population [14]. The social status and habits of the control patients were probably nearly the same as those of the THA patients; however, the THA patients' ability to move had been restricted long before the operation and remained less even after surgery. Their decreased mortality from accidents reflects their lower level of mobility than that of the control patients. All-cause mortality risk was significantly higher in physically inactive persons compared with active persons in a 5-year followup study [15]. Even in a shorter follow-up period and in a series of more than 10,000 THA patients, cardiovascular diseases were the cause of death in 59% within 3 months and in 37% from 3 to 12 months after the operation. During the first 3 postoperative months, the rate of the ischemic heart attack as a cause of death was 3.6-Iold compared with the rate during the following 9 months [16]. Regular recreational exercise habits were observed to increase substantially in part of this group after THA. The proportion of patients who engaged in regular walking increased from 2% to 55% because of THA [17]. The improved walking ability might stress the heart beyond its capabilities and result in sudden death. Careful initiation of physical exercises may be indicated after THA. The incidence of cardiovascular mortality in THA and control patients equalized after 16 years. The influence of the improved mobility of the THA patients in reducing cardiac deaths is probably slow. Although no antithrombosis prophylactic was given, there was only 1 death from thromboembolism in the present series in the early postoperative period. Three other cases were scattered through the whole follow-up period. The death rate of 0.39% from pulmonary embolism is low. Warwick et al. also report a low mortality rate of 0.34% trom pulmonary embolism during the 6 months after surgery in 1,162 consecutive THA patients. These patients also did not receive prophylactic anticoagulants [18]. Deep vein thrombo- sis is not uncommon after hospitalization, but very seldom does it have fatal consequences [19]. The distribution of the causes of death of THA patients did not differ much from that ot the general population in Finland. Only the cancer mortality was less than expected. Metallic ions and particles are released both from metal-on-metal and plastic-on-metal prostheses, resulting in a postoperative increase, especially of chromium, in serum levels and different organ levels [20-24]. Metal debris is widely disseminated in local and distant lymph nodes, bone marrow, spleen, and liver. It can cause necrosis of the lymph nodes, and an oncogenic effect is not excluded [6]; however, the THA and control patients had fewer cancer deaths than expected in the normal population. This is obviously due to the preoperative selection of these patients, because cancer patients are not usually operated on with elective orthopaedic procedures. In fact, the incidence of cancer in THA patients has been reported to be less than or equal to that of the general population [25-27]. Many forms of cancer are at present curable, so cancer mortality does not indicate the whole cancer risk after THA. The mean follow-up period of 12 years may be too short to express cancer mortality. A mean latent period of 25 years (range, 2-46 years) for local metallic foreign-body sarcoma development was required in war injuries [28]. This series shows that the THA patients have no increased risk of dying from cancer during the indicated follow-up period. References 1. Paavolainen E Slfitis P, Hfimfilfiinen Met al: Long-term results of total joint arthroplasty: results of a 15-year follow-up on a nationwide registration programme in Finland with TJAs. National Agency for Medicines, Medical Device Center, Helsinki, Laupacis A, Bourne R, Rorabeck C et al: The effect of elective total hip replacement on health-related quality of life. J Bone Joint Surg 75A:I619, Paffenberger RS, Hyde RT, Hsieh CC et al: Physical activity, other life-style patterns, cardiovascular disease and longevity. Acta Med Scand (suppl) 711:85, Pekkanen J, Marti B, Nissinen A et al: Reduction of premature mortality by high physical activity: a 20- year follow up of middle aged men. Lancet 1:1473, Powell KE, Thompson PD, Caspersen CJ et al: Physical activity and the incidence of coronary heart disease. Annu Rev Public Health 8:253, 1987

6 402 The Journal of Arthroplasty Vol. 12 No. 4 June Case CP, Langkamer VG, James C et al: Widespread dissemination of metal debris from implants. J Bone Joint Surg 76B:70I, Causes of death 199h Health 1992:8. Statistics Finland I Heli6vaara M, Reunanen A, Aromaa A et al: Validity of hospital data in a prospective epidemiological study on stroke and myocardial infarction. Acta Med Scand 216:309, Structure of population, p. 11. Central Statistical Office of Finland, Aro S, Koskinen R, Keskimfiki J: Sairaalapoistorekisterin diagnoosi-, toimenpide, ja tapaturmatietojen luotettavuus [Validity of the hospital discharge data]. Duodecim I06:1443, Visuri T, Pulkkinen P, Turula I(d3 et al: Life expectancy after hip arthroplasty: case-control study of 1018 cases of primary arthrosis. Acta Orthop Scand 65:9, Itolmberg S: Life expectancy after total hip arthroplasty. J Arthroplasty 7:183, Whittle J, Steinberg EE Anderson GF et al: Mortality after elective total hip arthroplasty in elderly Americans. Clin Orthop 295:119, Lindberg H, Nilsson BE: Coinciding morbidity in patients with coxarthrosis. Arch Orthop Trauma Surg 104:82, Wyerer S: Effects of physical inactivity of all-cause mortality risk in upper Bavaria. Perceptual Motor Skills 77:499, Seagrott V, Tan I-IS, Goldacre M et al: Elective total hip replacement: incidence, emergency readmission rate, and postoperative mortality. Br Med J 301:1431, Visuri T, Honkanen R: Total hip replacement: its influence on spontaneous recreation exercise habits. Arch Phys Med Rehab 325, Warwick D, Williams MH, Bannister GC: Death and thromboembolic disease after total hip replacement. J Bone Joint Surg 77B:6, Towbridge A, Boese CK, Wooruff Bet al: Incidence of posthospitalization proximal deep venous thrombosis after total hip arthroplasty. Clin Orthop 299:203, Bartolozzi A, Black J: Chromium concentrations in serum, blood clot and urine from patients following total hip arthroplasty. Biomaterials 6:2, Black J, Maitin EC, Gelman Het al: Serum concentrations of chromium, cobalt and nickel after total hip replacement: a six month study. Biomaterials 4:160, Langkamer VG, Case CP, Heap Pet ah Systemic distribution of wear debris after hip replacement. J Bone Joint Surg 74B:831, Michel R, Nolte M, Reich Met al: Systemic effects of implanted prostheses made of cobalt-chromium alloys. Arch Orthop Trauma Surg 110:61, Pazzaglia UE, Minoia C, Gulatieri Get al: Metal ions in body fluids after arthroplasty. Acta Orthop Scand 57:415, Gillespie WJ, Frampton CMA, Henderson RJ et al: The incidence of cancer following total hip replacement J Bone Joint Surg 70B:539, Mathiessen EB, Ahllbom A, Berman G et al: Total hip replacement and cancer. J Bone Joint Surg 77B:345, Visuri T, Koskenvuo M: Cancer incidence after McKee-Farrar total hip replacement. Orthopedics 14:137, Ott GH: Fremdk6rpersarkome. p. 1. Springer-Verlag, Berlin, 1970