 |
 |
 |
|
| home | CME | help | feedback | subscriptions | archive |
|
© 1998 The Journal of Bone and Joint Surgery, Inc. Postoperative Mortality after Total Hip Arthroplasty. An Analysis of Deaths after Two Thousand Seven Hundred and Thirty-six Procedures*  , BOSTON, MASSACHUSETTSInvestigation performed at the Orthopaedic Biomechanics Laboratory and the Hip and Implant Unit, Department of Orthopaedic Surgery, Massachusetts General Hospital, Boston
We retrospectively determined the prevalence and nature of mortality as many as ninety days after 2736 primary and revision total hip arthroplasties performed in 2002 patients by one surgeon at a teaching hospital between January 1969 and December 1996. All but seventy-one of the patients had received prophylaxis against venous thromboembolic disease. There were no intraoperative deaths, and no events during the operation could be linked directly to postoperative mortality. Eight deaths (mortality rate, 0.3 per cent) occurred within ninety days after the 2736 procedures. Four deaths (mortality rate, 0.15 per cent) occurred during the initial hospitalization. The cause of seven of the deaths was determined. Three patients died as a result of preexisting disease (severe hepatorenal disease, metastatic esophageal cancer, or severe cardiac disease), and one patient died from sepsis with a gram-negative organism during a thoracotomy eight days postoperatively. A bleeding complication that occurred while the patient was receiving warfarin therapy led to the death of two other patients; one of these deaths occurred in 1974 and the other, in 1982. At the time that these patients were managed, the desired prothrombin time was considered to be twice the control value. The remaining patient, who had had a clip placed on the inferior vena cava after a pulmonary embolus occurred in 1970, died secondary to acute, severe thrombosis of this vessel after a total hip arthroplasty in 1971. The patient for whom the cause of death was not determined had had an artificial aortic valve and had been receiving chronic warfarin therapy. She died suddenly eighty-nine days postoperatively; no autopsy was performed. No patient died as the direct result of a known pulmonary embolus. No deaths related to venous thromboembolic disease or its prophylaxis or treatment occurred after 1982 (1458 operations). We attribute this, in part, to reduced levels of warfarin prophylaxis and improved management with warfarin. The ninety-day postoperative mortality rate after 2736 procedures performed over nearly three decades was low (0.3 per cent). This span of time included the period before the introduction of many current improvements in perioperative care, such as routine intubation of patients under general anesthesia, continuous monitoring of the electrocardiogram intraoperatively, and blood-gas determinations. When the patients who died as a result of known, severe preexisting disease were excluded, the mortality rate was 0.18 per cent (five of 2733).
The purpose of the present study was to examine retrospectively the mortality rate during the first ninety days after 2736 consecutive total hip arthroplasties done in 2002 patients by one surgeon who specialized in this procedure at a major teaching institution. Such data can be valuable for patients who are considering total hip arthroplasty and can provide a basis for comparison with the experience of other surgeons and other institutions as well as with the results of other types of techniques. The ninety-day period was chosen to include all deaths that were likely to be related to the operation.
We retrospectively reviewed the records of all patients who had had a primary or revision total hip arthroplasty performed by the senior one of us (W. H. H.) between January 1969 and December 1996. Patients who had an intertrochanteric fracture, a fracture of the femoral neck, or a metastatic malignant tumor were included in the study. The fate of all patients for at least ninety days postoperatively was known. The records of patients who died within ninety days after the operation were studied in detail. Chi-square analysis was used to compare the mortality rates after primary and revision operations.
Study Population Prophylaxis against venous thromboembolic disease was used in conjunction with all but seventy-one total hip arthroplasties (seventy-one patients); fifty-one of the seventy-one patients were enrolled as the control group in an approved research study published in 1977, and the other twenty did not receive prophylaxis for various reasons. None of these seventy-one patients died. Warfarin was used in conjunction with 1794 procedures; aspirin, with 486; dextran, with 249; pneumatic compression boots, with ninety-two; and dextran in combination with pneumatic compression boots, with forty-four.
Eight deaths (mortality rate, 0.3 per cent) occurred within ninety days after the 2736 procedures. Four deaths (mortality rate, 0.15 per cent) occurred during the period of hospitalization for the initial operation, and four occurred after discharge. Four deaths (mortality rate, 0.2 per cent) occurred after the 2103 primary procedures and four (mortality rate, 0.6 per cent), after the 633 revision procedures; this difference was not found to be significant, with the numbers available (p = 0.07). Four patients had an autopsy in order to determine the cause of death; one of them died from a severe, chronic preexisting disease. In addition, two other patients died from a severe, chronic preexisting disease.
CASE 1. A sixty-eight-year-old woman with lupoid autoimmune hepatitis and cirrhosis had a revision total hip arthroplasty because of severe pain resulting from a fracture of the femoral stem. The postoperative course was uncomplicated except for a continued rise in the total bilirubin level from 2.3 to 5.4 milligrams per deciliter (39.3 to 92.3 micromoles per liter). (The upper limit of normal at our hospital is 1.0 milligram per deciliter [17.1 micromoles per liter].) The patient was doing well at the time of discharge to home, twenty-seven days postoperatively. She died from hepatorenal syndrome forty-nine days postoperatively. CASE 2. An eighty-two-year-old man with metastatic esophageal cancer, stable angina, and a history of a myocardial infarction involving the inferior wall had a total hip arthroplasty because of severe pain caused by avascular necrosis. The patient was discharged to home, but he died twenty-eight days postoperatively as a result of complications related to the esophageal cancer. CASE 3. An eighty-one-year-old man with a history of a myocardial infarction involving the inferior wall, severe aortic stenosis, and Paget disease had a so-called hybrid total hip arthroplasty (an acetabular component inserted without cement and a femoral component inserted with cement) to treat a severely painful non-union of a fracture of the femoral neck. The patient was discharged to home. Fifty-one days postoperatively, he sustained a fracture of the femur through pagetic bone distal to the tip of the prosthesis. Four days later, the patient died as a result of cardiac arrest as determined at autopsy. CASE 4. In 1970, acute tubular necrosis developed in a forty-one-year-old woman after a revision total hip arthroplasty that had been performed because a previous cup arthroplasty had failed. Eight days postoperatively, respiratory arrest developed and the patient was intubated. An emergency thoracotomy was performed because of a suspected massive pulmonary embolus. The patient died in the operating room. The autopsy revealed sepsis with a gram-negative organism, pulmonary edema, and severe acute tubular necrosis. CASE 5. In 1974, a seventy-seven-year-old woman had a revision total hip arthroplasty because of a failed hemiarthroplasty. The patient had had a previous infection with Staphylococcus epidermidis at the site of the hemiarthroplasty, but cultures were negative at the time of the revision. The patient was given warfarin with the aim of prolonging the prothrombin time to twenty-four seconds (twice the control value), which was considered the desired level at the time. On the ninth postoperative day, the patient had severe bleeding from the wound, hypotension, and cardiac arrest. Although the patient was resuscitated, she had acute renal failure and dehiscence of the wound. Her condition continued to worsen, and she died eighty-four days postoperatively. No autopsy was performed. CASE 6. A sixty-five-year-old man had a clip placed on the inferior vena cava because of a pulmonary embolus following a total hip arthroplasty in 1970. A total hip arthroplasty was performed on the contralateral hip in 1971; warfarin was used as a prophylaxis. Again, the desired prothrombin time at that time was considered to be twice the control value (twenty-four seconds). A retroperitoneal hematoma developed while the prothrombin time was twenty-six seconds, necessitating a laparotomy. Warfarin therapy was discontinued, and the patient was given vitamin K. Eight days later, severe deep-vein thrombosis developed and warfarin therapy was resumed. Thirty days postoperatively, the patient was found without a pulse and he died despite efforts at resuscitation. The autopsy revealed acute, occlusive thrombosis of the inferior vena cava and severe pulmonary congestion but no evidence of pulmonary emboli. CASE 7. A pulmonary embolus developed five days after a revision total hip replacement in a sixty-seven-year-old man who was taking aspirin as prophylaxis against deep-vein thrombosis. The embolus was treated with intravenous administration of heparin and warfarin. Bleeding from the wound necessitated arterial embolization on the thirteenth day postoperatively, but respiratory arrest developed on removal of the catheter from the femoral artery, resulting in brain damage. Life support in the form of mechanical ventilation was withdrawn on the twenty-second day postoperatively with the concurrence of the family. The autopsy confirmed that the cause of death was brain damage, not pulmonary embolus. CASE 8. A sixty-three-year-old woman who had hypertension and an artificial aortic valve was being managed with chronic warfarin therapy at the time of a primary hybrid total hip arthroplasty, which was performed without incident. The patient was doing well when she was seen in the physician's office seventy days postoperatively; however, she died suddenly on an airplane eighty-nine days postoperatively. No autopsy was performed.
Several features of this study are unique. First, all of the patients were managed by the same surgeon. Also, the series spans nearly thirty years of experience at a major teaching hospital and includes many patients who were managed operatively before the institution of several current improvements in preoperative, intraoperative, and postoperative care, such as routine intubation of patients under general anesthesia, discontinuation of the use of ether and cyclopropane as anesthetic agents, continuous monitoring of the electrocardiogram both intraoperatively and in the recovery room, blood-gas determinations, and the use of oximeters. In addition, patients who had a malignant tumor, an intertrochanteric fracture, or a fracture of the femoral neck were included in the study; the inclusion of such patients usually results in an increase in the mortality rate1,6,9. Finally, all deaths that occurred within ninety days after the operation were included in this study. Despite these features, the overall postoperative mortality rate was 0.3 per cent (eight deaths), with a rate of 0.2 per cent (four deaths) after the 2103 primary hip replacements and a rate of 0.6 per cent (four deaths) after the 633 revision hip replacements. These rates are lower than those recently reported in the literature2,4,5,7,8 (Table I).
Sharrock et al. reported a 0.1 per cent mortality rate in a series of 9685 patients who had had a primary hip replacement from 1987 to 1991 at The Hospital for Special Surgery. This rate was attributed in part to the use of regional anesthesia and low levels of induced hypotension7. Patients who had a fracture of the hip or a malignant tumor were excluded from the study, and only deaths that had occurred during the initial hospitalization were analyzed. In our study, the in-hospital mortality rate after primary total hip arthroplasty performed without the use of hypotensive, epidural anesthesia was the same as that reported by Sharrock et al.: 0.1 per cent (three of 2103) (Table I). The duration of follow-up is critical in a study of postoperative mortality. Ninety days appears to be appropriate, as advocated by Huber et al. For example, half of the deaths in our series occurred after discharge and five occurred thirty days or more postoperatively. Furthermore, the percentage of postoperative deaths occurring after discharge is likely to increase as the durations of hospital stays have become shorter. In a multicenter (156-site) trial comparing enoxaparin with warfarin for prophylaxis against deep-vein thrombosis, Colwell et al. reported a 0.74 per cent ninety-day mortality rate for 2155 patients who had had a total hip arthroplasty; they found that most of the deaths occurred after discharge. Taylor et al., in a nationwide study of 160,518 Medicare patients who had had a primary hip replacement during 1993 and 1994, found mortality rates of 0.74 per cent during the period of hospitalization and 1.42 per cent at thirty days. Although venous thromboembolic disease or attempts to prevent or treat it contributed to three deaths in this study, the overall mortality rate associated with venous thromboembolic disease and its prophylaxis was low (0.1 per cent). There were no known fatal pulmonary emboli. Wroblewski et al. reported a 0.7 per cent rate of fatal pulmonary emboli after 18,104 total hip procedures. The thromboembolism-related complications in this series would be less likely today for several reasons: aspirin is no longer used routinely for prophylaxis, inferior vena cava clips are no longer used, and vascular screening studies (such as venography and ultrasonography) allow early detection of deep-vein thrombosis. Perhaps more importantly, doses of warfarin have been decreased substantially, with a current desired prothrombin time of fifteen seconds (an international normalized ratio of 1.5) in our laboratory, with a control value of 12.2 seconds. No deaths related to the use of warfarin have occurred at our institution since 1982. In the present series, the volume of procedures performed by the surgeon may have been a factor, as noted in previous studies by Kreder et al. and by Lavernia and Guzman. Kreder et al. defined low-volume surgeons as those who are below the fortieth percentile with regard to volume (fewer than two primary hip replacements a year) and high-volume surgeons as those who are above the eightieth percentile (more than ten primary hip replacements a year). Lavernia and Guzman defined low volume as less than ten procedures a year and high volume as more than 100 procedures a year. Both groups of authors found a significantly lower ninety-day mortality rate after total hip replacements performed by high-volume surgeons than after replacements performed by low-volume surgeons (p < 0.05 in the study by Kreder et al. and p < 0.003 in the study by Lavernia and Guzman). The limitations of the present study include its retrospective nature; the somewhat younger average age of the patients compared with that in other studies; variations in care in general, and in anesthetic techniques specifically, over the past three decades; and the absence of complete data on comorbidities. The strength of this study is that the fate of every patient during the first ninety postoperative days was known. The mortality rate was low (0.3 per cent; eight deaths after 2736 procedures), and it was particularly low (0.18 per cent; five deaths after 2733 procedures) when the three patients who had a known severe coexisting disease are excluded.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was the William H. Harris Foundation, Boston, Massachusetts.
  CiteULike  Connotea  Del.icio.us  Technorati What's this?
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Top
Abstract
Introduction
| home | CME | help | feedback | subscriptions | archive | site index |
|
Stanford University Libraries' HighWire Press (TM) assists in the publication of JBJS Online. |