THA is associated with improved function and reduction in the risk of mortality, with numerous studies showing that patients having primary THA have a survival advantage relative to the general population during the first 5 to 8 postoperative years [1, 13, 14, 16, 18-20, 22, 24, 25]. Since the majority of primary THAs are elective, and patients typically are screened preoperatively for comorbid risks, the survival advantage likely represents selection of healthier individuals for surgery. Additionally, rates of utilization for THA vary across ethnic and socioeconomic groups, which also might contribute to this observed survival advantage [2, 8, 11, 21]. Nevertheless, one study that controlled for age, sex, race, comorbid diagnoses, and Medicare qualification found persistence of a survival advantage among patients undergoing primary THA, suggesting that factors such as postoperative improvement in functional status, decreases in use of pain medication, and perhaps the surgery itself also may offer some contribution .
It is unclear whether the same survival advantage applies to revision THA, a procedure for which the demand is expected to rise [4, 12, 15, 23]. A better understanding of mortality after revision THA is thus an important priority. Mortality studies to date after revision THA have largely focused on short-term mortality . For example, a recent study from the Danish Hip Arthroplasty Register found a twofold higher risk of death within the first year after revisions for periprosthetic joint infections . Numerous factors hinder the study of longer-term outcomes in the revision THA populations, particularly lack of large, well-documented cohorts with long-term followup data. Using an institutional joint registry, we therefore investigated survivorship trends in a large cohort of patients who underwent revision THA.
Our specific goals were to examine (1) long-term mortality according to reasons for revision after revision THA, and (2) relative mortality trends by age at surgery, years since surgery, and calendar year of surgery.
Study Design and Setting
We performed a retrospective study using data from our institutional joint registry from 1969 to 2011. The study was approved by the Mayo Clinic institutional review board. Patients who had declined research authorization for use of their medical records in research had less than 5% of procedures and were excluded from the study. There were no other exclusion criteria. The final study cohort comprised 5417 revision THAs in 4532 patients between January 1, 1971 and December 31, 2011. The 5417 revision THAs included first time and subsequent revisions. December 2011 was chosen as the end date to allow sufficient followup of the study cohort. Of the 4532 patients, 3820 (84%) had only one revision THA at our institution. The remaining 712 patients had two or more revision THAs, regardless of laterality (ie, 574 had two, 112 had three, 23 had four, three had five, and one had six revision THAs).
Description of Experiment, Treatment, or Surgery
Demographic and THA-related clinical data were obtained from the joint registry and associated patient records. Data elements included comorbidities, clinical and operative details, including surgical indications for revision and operative date. Comorbidities were ascertained for each patient at the time of revision surgery using an electronic adaptation of the comorbidity index developed by Charlson et al. [5, 6]. Our institution has a long history of coding medical and surgical diagnoses starting with the in-house Berkson codes between 1966 and 1975 and Hospital Adaptation of the International Classification of Diseases (H-ICDA) codes between 1976 and 2010. Between 1966 and 2010, codes were assigned manually by trained nosologists through manual review of the medical and surgical records. Billing data-based ICD-9 coding started in 1995. As part of the Mayo Clinic research infrastructure, we have access to validated tools that provide a cross-walk between Berkson-H-ICDA–ICD 9 codes and allow classification of comorbidities according to the Charlson Comorbidity Index.
Patients who had revision THA were grouped by surgical indication in three categories: aseptic loosening, bearing wear, or dislocation; periprosthetic joint infection; and periprosthetic fracture. The indications for the 5417 revision THAs were 3833 (71%) revisions for loosening, wear, or dislocation; 938 (17%) revisions for periprosthetic joint infections; and 646 (12%) revisions for periprosthetic fractures. In the 3833 revisions for loosening, wear, or dislocation, there were 97 (2.5%) revisions with joint instability as the main surgical indication. Given the small size of this subset, patients with instability or dislocation were combined with patients with other aseptic, atraumatic indications of aseptic loosening and bearing wear. The outcome of interest was death, which is an outcome at the patient level. Therefore, individuals with multiple revised THAs (regardless of laterality) were counted only once, and surgical indications were classified hierarchically. In individuals with multiple revision THAs, if all of the revision procedures were for aseptic loosening, wear, or dislocation, then that patient was classified as having aseptic loosening, wear, or dislocation. If at least one of the revision surgeries was for fracture or periprosthetic joint infection, then that patient was classified as having fracture or infection, respectively. Patients who had revisions for infection and periprosthetic fracture were classified as having periprosthetic joint infection. This resulted in 3045 patients classified as having aseptic loosening or wear, 592 classified as having fractures, and 895 classified as having periprosthetic joint infections (total 4532 patients). The mean age of the entire cohort was 65 ± 13 years and 55% were males.
Patients were followed (either through a visit at the Mayo Clinic or by mail communication from local orthopaedists) at regular intervals at least twice during the first postoperative year, at 2, 5, 7, and 10 years, and at 5-year intervals thereafter until death or December 31, 2016. Deaths were recorded regardless of joint-specific registry followup. Mean followup of the entire cohort was 10.2 years (range, 1 day-37.2 years). At the time of the study, more than half of the cohort were deceased (2565; 57%) and 1967 (43%) were still alive and under observation.
Statistical Analysis, Study Size
Statistical analyses were performed for the cohort that had revision THAs as a whole and for each surgical indication category. Mortality rates were evaluated using a person-years approach, in which the observed number of deaths in the revision THA cohort was compared with the expected number of deaths during the followup period. This approach also was applied to the surgical indication categories. Because 97% of patients with recorded race data were white, life tables for the United States white population were used to generate the expected survival. The expected number of deaths was calculated assuming that the study cohort had the same calendar year-, age-, and sex-specific mortality rates as the United States general white population. The expected number of deaths was calculated for each calendar year, age, and sex strata and then summed to derive the total expected number of deaths. The ratio of observed to expected number of deaths was expressed as the standardized mortality ratio (SMR). SMRs were calculated for the entire revision THA cohort and for each surgical indication category. The CIs for the SMRs were calculated assuming that expected mortality rates were fixed and that observed deaths followed a Poisson distribution. Although the analysis was conducted separately for men and women, there were no sex-specific significant differences. Therefore, results are presented for men and women combined. In the calculations of the SMR for temporal trends, only events and person-years of exposure during the followup window of interest were considered. Poisson regression models were used to model relative mortality by age, time since surgery, and calendar year of surgery. Kaplan-Meier curves for survival after surgery were adjusted for age and sex. Cox proportional hazards regression was used to evaluate the association of survival and potential risk factors, including the number of revision surgeries. All analyses were performed using SAS Version 9.4 (SAS Institute Inc, Cary, NC, USA) and R version 3.2.0 (R Core Team, R Foundation for Statistical Computing, Vienna, Austria).
Long-term Mortality Risk By Indication After Revision THA
During a mean 10.2 ± 7.0 years followup of the entire cohort, we identified 2565 deaths. Comparison of the age- and sex-adjusted death rates in the three groups indicated that the highest age- and sex-adjusted death rate in the first year was among patients with periprosthetic fractures, with 2.0% deceased at 6 months and 3.5% at 12 months (Table 1). In contrast, among patients with aseptic loosening, wear, or dislocation-related THA revisions, the death rate was 0.7% at 6 months and 1.5% at 12 months. Among patients with periprosthetic joint infection-related revision THA, although the death rate within the first 6 months was twice as high as for patients with aseptic revisions (1.4% at 6 months), it remained lower than the rate for the fracture group. By 5 years, the age- and sex-adjusted death rate among patients with periprosthetic joint infections (16.0%) reached the same magnitude as for patients with periprosthetic fractures (17.5%). Beyond 10 years, the death rate remained highest among patients with periprosthetic joint infections (65.5% at 15 years and 81% at 20 years). The death rate among patients with aseptic loosening, wear, or dislocation caught up with the other two groups beyond 15 years and 75% of these patients were deceased at 20 years.
Compared with individuals in the general population, the overall mortality rate of patients with revision THAs was significantly higher (SMR, 1.09; 95% CI, 1.05-1.13; p < 0.001) (Table 2). There were notable differences across the three surgical indication subcategories. The relative mortality rate was similar to mortality in the general population among patients who underwent revision THA for aseptic loosening, wear, or dislocation (SMR, 1.01; 95% CI, 0.96-1.06; p = 0.647). The relative mortality rate was significantly higher than that of the general population among patients who underwent revision THA for periprosthetic fractures (SMR, 1.23; 95% CI, 1.11-1.37; p < 0.001) and for periprosthetic joint infections (SMR, 1.35; 95% CI, 1.24-1.48; p < 0.001).
In terms of comorbidities, the mean (SD) Charlson score at the time of surgery was 2.96 (± 1.93) in the aseptic loosening, wear, or dislocation cohort, 3.22 (± 2.34) in the periprosthetic fractures cohort, and 2.99 (± 2.16) in the periprosthetic joint infections cohort. Although not statistically significant (p > 0.05), patients with periprosthetic joint infections had the highest prevalence of diabetes with or without organ damage (15% versus 11% in patients with fractures and 10% in patients with aseptic loosening, wear, or dislocation), chronic pulmonary diseases (12% versus 10% in patients with fractures and 8% in patients with aseptic loosening, wear, or dislocation), rheumatologic diseases (9% versus 7% in patients with fractures and 6% in patients with aseptic loosening, wear, or dislocation), and cancer (9% versus 7% in patients with fractures and 8% in patients with aseptic loosening, wear, or dislocation).
Relative Mortality Trends
The relative mortality trends are provided by age at surgery (Fig. 1) (see Appendix, Supplemental Digital Content 1), years since surgery (Fig. 2) (see Appendix, Supplemental Digital Content 2, and calendar year of surgery (Fig. 3) (see Appendix, Supplemental Digital Content 3). Relative mortality rates were higher than for the general population in younger age groups until approximately 60 to70 years of age and declined with increasing age at surgery (Fig. 1). This pattern was observed in all three indication groups. For example, among patients who underwent revision THA for aseptic loosening and/or wear, the relative mortality rate remained higher than that for the general population until approximately 70 years of age, with an SMR of 1.09 (95% CI, 1.04-1.14) at 70 years. Among patients with periprosthetic joint infection, the relative mortality rates remained significantly higher than those for the general population throughout the entire age spectrum as evidenced by confidence intervals not including 1.00 except in the very elderly.
The mortality rate during the first year of surgery was lower than that of the general population among patients with aseptic and atraumatic revisions (Fig. 2). However, the relative mortality rate of these patients increased with time and we observed a shift at approximately 8 to 10 years after surgery. The SMR was 0.85 (95% CI, 0.82-0.88) at 5 years and 1.19 (95% CI, 1.14-1.23) at 10 years. The relative mortality rate of patients who had revisions for infections and fractures was higher than that for patients with aseptic and atraumatic revisions throughout entire time. We also observed an early rise in mortality in the fracture and periprosthetic joint infection groups with SMRs of 1.16 (95% CI, 1.10-1.23) and 1.25 (95% CI, 1.19-1.31) respectively at 5 years.
The relative mortality rates according to calendar year of surgery also differed across the three groups (Fig. 3). Most notably, the relative mortality rate in the aseptic loosening revision group showed almost a linear decline with time, indicating steady mortality improvements with time. We did not observe a similar improvement in mortality for the other two groups. Although the visual inspection of the relative mortality in the fracture and infection groups appear to have worsened with time, this was not significant (p < 0.05).
Finally, we found no evidence that the number of revisions is a significant predictor of mortality (p = 0.662) in an age-, sex-, and calendar year-adjusted Cox model.
The survival advantage after THA is well documented. This is attributed primarily to preoperative screening and selection of healthier individuals for THA. However, it is unknown whether the survival advantage persists after revision THA. We investigated long-term mortality trends in a cohort of 4532 patients who had undergone revision THA during the last 5 decades at a tertiary care hospital in the United States. Our study indicates that patients who undergo revision THA for aseptic indications experience a lower risk of death compared with the general population during the first year after surgery, but the risk increased with time and got worse than that of the general population after 8 to 10 years after surgery. However, patients who undergo revision THA for infections and fractures experience a significant excess risk of death during the first year of surgery that persists several years after surgery. Our study also provides evidence for secular improvements in mortality after revision THA for aseptic indications, but not for infections and fractures. These findings reinforce the need for long-term followup of patients undergoing THA, not only for implant survival but overall health.
Our study has several potential limitations. First, although we accounted for all revision THAs at our institution, this is an incomplete account of the total number of procedures for each patient. Patients may have had revisions before and/or after receiving care at our institution (on the same joint or the contralateral joint). Therefore, the lack of a significant association with the number of revisions in this cohort should be interpreted with caution. Second, any categorization of patients by surgical indication must deal with numerous challenges. For example, many patients have multiple procedures and some may have revisions for multiple indications. Patients may have one revision for aseptic loosening and a subsequent revision for infection on the contralateral joint. Despite these limitations, our overall findings suggest that the surgical indication may be a more-important driver of revision THA mortality than the number of revision surgeries. Third, the cohort was assembled retrospectively at a large tertiary care hospital and may not be representative of all patients having revision THA in the United States owing to differences in ethnicity, socioeconomic status, and healthcare access. The generalizability of our findings to ethnically, racially, and socioeconomically more-diverse populations is unknown. Fourth, this study extends over four decades and THA surgical practice evolved with time. Surgeons in the current era may be more aggressive than in the past in operating on patients with multiple comorbid conditions. Finally, we were unable to compare our patient comorbidity data with that of the general population as it was not feasible to obtain national mortality data by comorbidity groups. In a long-term mortality study, comorbidities at the time of surgery are an incomplete account of the lifetime comorbidity burden in these patients.
An early survival advantage was confirmed in our revision THA cohort for patients who underwent surgery for aseptic indications other than periprosthetic fractures. A previous investigation showed a mean time to revision THA of approximately 7 years after a primary THA . Given that previous studies of primary THA have found a postoperative survival advantage of 5 to 8 years [1, 10, 13, 14, 16, 18-20, 22, 24, 25], many patients in our revision THA cohort may have already passed beyond the window in which an additional survival advantage would be observed. Rather, our findings suggest a high likelihood of patient selection at the time of revision surgery, similar to the observed effect in primary THA. Alternatively, improved functional status may contribute to the decreased mortality risk in patients undergoing revision THA for elective aseptic loosening, wear, or dislocation . In contrast, patients with periprosthetic joint infection and fracture have not experienced the same selection advantage, as they often are emergent rather than elective surgeries with longer operative times and multiple stages. Further studies are warranted to understand whether the observed mortality patterns are purely attributable to selection bias or if other factors such as comorbidities, surgery effects, and healthcare utilization play an important role.
Notably, patients who undergo revision THA for periprosthetic joint infection have the highest mortality risk and the excess risk persists for years after surgery. This pattern is consistent with that reported in a previous study , and likely is attributable to associated comorbidities and risk factors in patients with periprosthetic joint infection. For instance, obesity is a risk factor for periprosthetic joint infections, and the risk of infection increases notably with increasing BMI [7, 17, 26]. Other associated risk factors for periprosthetic joint infection such as diabetes mellitus, cardiovascular disorders, and substance abuse may contribute to excess mortality [7, 17]. Our comorbidity analyses confirm the higher comorbidity burden in patients with periprosthetic joint infections, with 15% prevalence of diabetes, 12% chronic pulmonary disease, and 9% each of rheumatologic diseases and cancer. Consequently, the excess mortality rates among patients with periprosthetic joint infection are likely attributable to their overall level of health in addition to the revision procedure.
Our examination of mortality trends by age showed that relative mortality is high in younger patients. This is similar to what has been reported in patients having primary THA . Patients undergoing surgery at a younger age can have important comorbidities, such as inflammatory arthritis, and therefore may have a higher mortality risk. Second, relative mortality trends by time since surgery indicate that relative mortality is approximately 50% lower than that of the general population among patients with aseptic revision surgeries, and that this survival advantage persists for at least 8 years after surgery. This pattern also been has reported for patients having primary THA . However, for patients who underwent surgery for infection, the relative mortality rate is high even during the first year after surgery. Third, we observed steady declines in mortality risk during the last 2 decades after surgeries for aseptic indications. Again, this is similar to the secular mortality declines previously reported following primary THA [10, 16], and attests to improvements in surgical techniques and perioperative care. Lack of a similar secular decline for infection and fracture revision indications is possibly attributable to their nonelective nature, the extent of the surgical procedures, or a combination of these and other undetermined factors
Secular declines in mortality risk for the majority of patients having revision THA are reassuring. However, patients undergoing revision THA for periprosthetic joint infections continue to experience a higher mortality risk even several years after surgery. Shifting mortality patterns several years after surgery reinforce the need for long-term followup, not only for implant survival, but for overall health of patients having THA.
1. Barrett J, Losina E, Baron JA, Mahomed NN, Wright J, Katz JN. Survival following total hip replacement. J Bone Joint Surg Am. 2005;87:1965–1971.
2. Bass AR, McHugh K, Fields K, Goto R, Parks ML, Goodman SM. Higher total knee arthroplasty revision rates among United States blacks than whites: a systematic literature review and meta-analysis. J Bone Joint Surg Am. 2016;98:2103–2108.
3. Berstock JR, Beswick AD, Lenguerrand E, Whitehouse MR, Blom AW. Mortality after total hip replacement surgery: a systematic review. Bone Joint Res. 2014;3:175–182.
4. Bozic KJ, Kurtz SM, Lau E, Ong K, Vail TP, Berry DJ. The epidemiology of revision total hip arthroplasty in the United States. J Bone Joint Surg Am. 2009;91:128–133.
5. Charlson ME, Pompei P, Ales KL, MacKenzie CR. A new method of classifying prognostic comorbidity in longitudinal studies: development and validation. J Chronic Dis. 1987;40:373–383.
6. Deyo RA, Cherkin DC, Ciol MA. Adapting a clinical comorbidity index for use with ICD-9-CM administrative databases. J Clin Epidemiol. 1992;45:613–619.
7. Eka A, Chen AF. Patient-related medical risk factors for periprosthetic joint infection of the hip and knee. Ann Transl Med. 2015;3:233.
8. Escalante A, Espinosa-Morales R, del Rincon I, Arroyo RA, Older SA. Recipients of hip replacement for arthritis are less likely to be Hispanic, independent of access to health care and socioeconomic status. Arthritis Rheum. 2000;43:390–399.
9. Gundtoft PH, Pedersen AB, Varnum C, Overgaard S. Increased mortality after prosthetic joint infection in primary THA. Clin Orthop Relat Res. 2017 . [Epub ahead of print] doi: .
10. Hunt LP, Ben-Shlomo Y, Clark EM, Dieppe P, Judge A, MacGregor AJ, Tobias JH, Vernon K, Blom AW; National Joint Registry for England, Wales, and Northern Ireland. 90-day mortality after 409,096 total hip replacements for osteoarthritis, from the National Joint Registry for England and Wales: a retrospective analysis. Lancet. 2013;382:1097–1104.
11. Ibrahim SA, Kim H, McConnell KJ. The CMS comprehensive care model and racial disparity in joint replacement. JAMA. 2016;316:1258–1259.
12. Kurtz SM, Lau E, Ong K, Zhao K, Kelly M, Bozic KJ. Future young patient demand for primary and revision joint replacement: national projections from 2010 to 2030. Clin Orthop Relat Res. 2009;467:2606–2612.
13. Lie SA, Engesaeter LB, Havelin LI, Gjessing HK, Vollset SE. Mortality after total hip replacement: 0-10-year follow-up of 39,543 patients in the Norwegian Arthroplasty Register. Acta Orthop Scand. 2000;71:19–27.
14. Lindberg H, Carlsson AS, Lanke J, Horstmann V. The overall mortality rate in patients with total hip arthroplasty, with special reference to coxarthrosis. Clin Orthop Relat Res. 1984;191:116–120.
15. Maradit Kremers H, Larson DR, Crowson CS, Kremers WK, Washington RE, Steiner CA, Jiranek WA, Berry DJ. Prevalence of total hip and knee replacement in the United States. J Bone Joint Surg Am. 2015;97:1386–1397.
16. Maradit Kremers H, Larson DR, Noureldin M, Schleck CD, Jiranek WA, Berry DJ. Long-term mortality trends after total hip and knee arthroplasties: a population-based study. J Arthroplasty. 2016;31:1163–1169.
17. Maradit Kremers H, Lewallen LW, Lahr BD, Mabry TM, Steckelberg JM, Berry DJ, Hanssen AD, Berbari EF, Osmon DR. Do claims-based comorbidities adequately capture case mix for surgical site infections? Clin Orthop Relat Res. 2015;473:1777–1786.
18. Paavolainen P, Pukkala E, Pulkkinen P, Visuri T. Causes of death after total hip arthroplasty: a nationwide cohort study with 24,638 patients. J Arthroplasty. 2002;17:274–281.
19. Pedersen AB, Baron JA, Overgaard S, Johnsen SP. Short- and long-term mortality following primary total hip replacement for osteoarthritis: a Danish nationwide epidemiological study. J Bone Joint Surg Br. 2011;93:172–177.
20. Ramiah RD, Ashmore AM, Whitley E, Bannister GC. Ten-year life expectancy after primary total hip replacement. J Bone Joint Surg Br. 2007;89:1299–1302.
21. Ravi B, Croxford R, Austin PC, Lipscombe L, Bierman AS, Harvey PJ, Hawker GA. The relation between total joint arthroplasty and risk for serious cardiovascular events in patients with moderate-severe osteoarthritis: propensity score matched landmark analysis. BMJ. 2013;347:f6187.
22. Surin VV, Sundholm K. Survival of patients and prostheses after total hip arthroplasty. Clin Orthop Relat Res. 1983;177:148–153.
23. Ulrich SD, Seyler TM, Bennett D, Delanois RE, Saleh KJ, Thongtrangan I, Kuskowski M, Cheng EY, Sharkey PF, Parvizi J, Stiehl JB, Mont MA. Total hip arthroplasties: what are the reasons for revision? Int Orthop. 2008;32:597–604.
24. Visuri T, Pulkkinen P, Paavolainen P, Koskenvuo M, Turula KB. Causes of death after hip arthroplasty in primary arthrosis. J Arthroplasty. 1997;12:397–402.
25. Visuri T, Pulkkinen P, Turula KB, Paavolainen P, Koskenvuo M. Life expectancy after hip arthroplasty: case-control study of 1018 cases of primary arthrosis. Acta Orthop Scand. 1994;65:9–11.
26. Wagner ER, Kamath AF, Fruth KM, Harmsen WS, Berry DJ. Effect of body mass index on complications and reoperations after total hip arthroplasty. J Bone Joint Surg Am. 2016;98:169–179.