Models adjusted for age, race, ethnicity, employment status, Medicaid dual eligibility, cause of end-stage renal disease, dialysis modality, dialysis vintage, congestive heart failure, coronary artery disease, other heart disease, cerebrovascular disease, peripheral arterial disease, hypertension, diabetes, and current smoking. Error bars indicate 95% CIs.
The adjusted models were adjusted for age, race, ethnicity, employment status, Medicaid dual eligibility, cause of end-stage renal disease, dialysis modality, dialysis vintage, congestive heart failure, coronary artery disease, other heart disease, cerebrovascular disease, peripheral arterial disease, hypertension, diabetes, and current smoking.
eTable 1. Changes in Comorbidity Labels in the 1995 and 2005 Versions of the Medical Evidence Report
eTable 2.International Classification of Diseases, Ninth Edition (ICD-9) procedure codes and Current Procedural Terminology, 4th Edition (CPT-4) used to Identify Lower Extremity Amputation, to Exclude Suspected Nonatherosclerotic Causes of Amputation, and to Identify Endovascular and Surgical Revascularization Procedures
eTable 3. Adjusted Absolute Rate (per 100 Person-years) and Rate Ratio (Reference, 2000) of Lower Extremity Amputations by Cohort Year
eTable 4. Adjusted Absolute Rate (per 100 Person-years) and Rate Ratio (Reference, 2000) of Lower Extremity Revascularizations by Cohort Year
eTable 5. Adjusted Absolute Rate (per 100 Person-years) and Rate Ratio (Reference, 2000) of Lower Extremity Amputations by Cohort Year, for Patients With Diabetes, Without Diabetes, and Without Diabetes or Hypertension
eTable 6. Adjusted Absolute Rate (per 100 Person-years) and Rate Ratio (Reference, 2000) of Lower Extremity Amputations by Cohort Year, Stratified by Age and Sex
eFigure. Adjusted Trends in the Rates of Lower Extremity Revascularization Procedures in Patients Receiving Dialysis From 2000 Through 2014
Customize your JAMA Network experience by selecting one or more topics from the list below.
Identify all potential conflicts of interest that might be relevant to your comment.
Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.
Err on the side of full disclosure.
If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.
Not all submitted comments are published. Please see our commenting policy for details.
Franz D, Zheng Y, Leeper NJ, Chandra V, Montez-Rath M, Chang TI. Trends in Rates of Lower Extremity Amputation Among Patients With End-stage Renal Disease Who Receive Dialysis. JAMA Intern Med. 2018;178(8):1025–1032. doi:10.1001/jamainternmed.2018.2436
Have rates of lower extremity amputation among patients with end-stage renal disease who receive dialysis changed in a recent 15-year period?
This study assessing more than 3 million records in a US national end-stage renal disease registry found that rates of lower extremity amputation decreased from 5.4 to 2.7 per 100 person-years from 2000 to 2014. Among patients who had an amputation, nearly half died the following year.
Despite improvements, patients with end-stage renal disease who receive dialysis continue to display high rates of lower extremity amputation and postamputation 1-year mortality.
Patients with end-stage renal disease (ESRD) who receive dialysis are at high risk of lower extremity amputation. Recent studies indicate decreasing rates of lower extremity amputation in non-ESRD populations, but contemporary data for patients with ESRD who receive dialysis are lacking.
To assess rates of lower extremity amputation among patients with ESRD who receive dialysis during a recent 15-year period; to analyze whether those rates differed by age, sex, diabetes, or geographic region; and to determine 1-year mortality rates in this population after lower extremity amputation.
Design, Setting, and Participants
This retrospective study of 3 700 902 records obtained from a US national registry of patients with ESRD who receive dialysis assessed cross-sectional cohorts for each calendar year from 2000 through 2014. Adult patients with prevalent ESRD treated with hemodialysis or peritoneal dialysis covered by Medicare Part A and B on January 1 of each cohort year were included. Data analysis was conducted from August 2017 to April 2018.
Age, sex, diabetes, and hospital referral region.
Main Outcomes and Measures
Annual rates per 100 person-years of nontraumatic major (above- or below-knee) and minor (below-ankle) amputations.
For each annual cohort, there were fewer women (47.5% in 2000, 46.2% in 2005, 44.9% in 2010, and 44.0% in 2014) than men, more than half the patients were white individuals (58.1% in 2000, 56.9% in 2005, 56.9% in 2010, and 56.7% in 2014), and a small proportion were employed (13.9% in 2000, 15.1% in 2005, 16.1% in 2010, and 16.5% in 2014). The rate of lower extremity amputations for patients with ESRD who receive dialysis decreased by 51.0% from 2000 to 2014, driven primarily by a decrease in the rate of major amputations (5.42 [95% CI, 5.28-5.56] in 2000 vs 2.66 [95% CI, 2.59-2.72] per 100 person-years in 2014). Patients with diabetes had amputation rates more than 5 times as high as patients without diabetes. Patients younger than 65 years had higher adjusted amputation rates than older patients, and men had consistently higher adjusted amputation rates than women. Adjusted 1-year mortality rates after lower extremity amputation for patients with ESRD who receive dialysis decreased from 52.2% (95% CI, 50.9%-53.4%) in 2000 to 43.6% (95% CI, 42.5%-44.8%) in 2013. In general, amputation rates decreased among all regions from 2000 to 2014, but regional variability persisted across time despite adjustment for differences in patient demographics and comorbid conditions.
Conclusions and Relevance
Although rates of lower extremity amputations among US patients with ESRD who receive dialysis decreased by 51% during a recent 15-year period, mortality rates remained high, with nearly half of patients dying within a year after lower extremity amputation. Our results highlight the need for more research on ways to prevent lower extremity amputation in this extremely high-risk population.
Peripheral arterial disease (PAD) affects approximately 8.5 million persons in the United States, with an annual incidence rate of 2.8% and prevalence rate of 12.3%.1 Patients with the most severe forms of PAD require lower extremity amputation, and the aftermath can be devastating.2 Patients with end-stage renal disease (ESRD) who receive dialysis are at especially high risk of amputation,3-7 perhaps due in part to the high prevalence of traditional risk factors, such as diabetes and hypertension, and of ESRD-specific risk factors, such as chronic inflammation and uremia.8-10 Initiation of dialysis is also an independent risk factor for amputation.5-7
An analysis of the general Medicare population showed a 45% decrease in lower extremity amputation rates from 1996 to 2011.11 Commonly cited reasons for this improvement include better PAD screening and vascular care, especially for patients with diabetes.2,11 Whether there has been a similar improvement in rates of lower extremity amputation for patients with ESRD who receive dialysis is unknown because the rates in this population were last described nearly 2 decades ago.12
Thus, we used the national ESRD registry to evaluate the 2000 to 2014 rates of lower extremity amputation for patients with ESRD who receive dialysis and whether those rates were associated with patient characteristics or comorbidities. We also examined regional variation in amputation practice patterns and 1-year mortality rates after lower extremity amputation.
We used the US Renal Data System (USRDS), a national data system that collects, analyzes, and distributes information about ESRD in the United States.13 Data were gathered from the USRDS for all patients initiating hemodialysis or peritoneal dialysis between January 1, 1996, and October 1, 2014. We generated annual cross-sectional populations of patients with prevalent ESRD who receive dialysis and are covered by Medicare (Parts A and B) on the first day of January for each calendar year from 2000 to 2014, creating 15 annual cohorts. We limited the cohorts to patients initiating dialysis on or after January 1, 1996, because the Medical Evidence Report form CMS 2728 that we used to ascertain comorbid conditions is completed within 45 days of incident ESRD, but this form was not required until 1995. Owing to the accrual of prevalent patients with time, the size of each annual cohort progressively increased, as did the time after initiation of dialysis (ie, dialysis vintage; Table 1). Patients were excluded if they were younger than 18 years or older than 110 years of age at incident ESRD, if their CMS 2728 form was missing or filed more than 45 days after initiation of dialysis, or if data for sex or race were missing. We also excluded patients who recovered kidney function within 365 days of dialysis initiation. The institutional review board of Stanford University School of Medicine (Palo Alto, California) approved this study and deemed that the need for patient consent could be waived because the data were deidentified and a waiver would not adversely affect the rights and welfare of the participants.
We collected information from the USRDS patient and treatment history files on age, sex, race (white, black, Asian, Native American, or other), Hispanic ethnicity, Medicaid eligibility, employment status, presumed cause of ESRD (diabetes, hypertension, glomerulonephritis, or other), dialysis modality, dialysis vintage, and year of dialysis initiation. Information was collected from the Medical Evidence Report on comorbid conditions, which were selected a priori given their associations with lower extremity amputations in ESRD and included PAD, heart failure, coronary artery disease, diabetes, hypertension, other heart disease (including arrhythmias and pericarditis), cerebrovascular disease, and current smoking.8 Some of the comorbidity category names in the Medical Evidence Report changed from the 1995 version to the 2005 version and were thus renamed as necessary (eTable 1 in the Supplement). For example, cardiac failure on the 1995 form was changed to congestive heart failure on the 2005 form, and both were categorized as heart failure in our analysis. We also collected residence location by hospital referral region on January 1 of each cohort year to evaluate the regional variation in amputation rates.14
The primary outcome of interest was the number of lower extremity amputations per 100 patient-years for each cohort year, identified using International Classification of Diseases, Ninth Edition (ICD-9) procedure codes and Current Procedural Terminology, 4th Revision (CPT-4) codes (eTable 2 in the Supplement). Amputations were categorized as major (above- or below-knee) or minor (below-ankle). Amputations occurring during the same encounter as one with diagnosis codes for trauma, congenital or acquired deformities, or malignant neoplasms were excluded. To prevent amputation event rate overestimation due to planned multiple-step procedures that may occur across weeks or months, we included only the highest-level amputation per patient per calendar year, as has been previously done.15 Thus, our results for the number of amputations per year were likely a conservative estimate.
We determined the number of patients who had a lower extremity endovascular revascularization, a surgical bypass, or both for each cohort year using ICD-9 and CPT-4 billing codes (eTable 2 in the Supplement).16-18
The second outcome of interest was the 1-year mortality rate after lower extremity amputation. This outcome applied only to those patients within each annual cohort who had undergone amputation in 2000 through 2013 to enable a 1-year follow-up for all cohorts. One-year mortality was defined as death within 365 days of the first date of the highest-level amputation for the cohort year.
To assess whether the rate of lower extremity amputation changed with time, we used a log Poisson generalized linear model with robust standard errors to model the number of patients with amputations per 100 person-years for each cohort year. Exposure time began on January 1 of the cohort year and ended with the first of the following events: death, loss of Medicare dual eligibility, or December 31 of cohort year. Models included an offset term with log exposure time. We adjusted for the following variables: age, race, ethnicity, employment status, Medicaid dual eligibility, cause of ESRD, dialysis modality, dialysis vintage, congestive heart failure, coronary artery disease, other heart disease, cerebrovascular disease, PAD, hypertension, diabetes, and current smoking.8 We then calculated adjusted rates, rate ratios, and their 95% CIs for each cohort year using the first cohort year (2000) as a reference. Unadjusted and adjusted amputation rates were calculated overall and by level of amputation. We also examined differences in amputation rates by diabetes status (defined as having diabetes as a cause of ESRD, a comorbid condition, or both), age (<65 years or ≥65 years), and sex. Amputation rates among patients without diabetes or hypertension (defined as having hypertension as a cause of ESRD, a comorbid condition, or both) were also assessed. We tested for effect modification of the linear trend with time by diabetes, age, or sex using a linear regression model applied directly to the estimated rates.
To evaluate regional variation, we examined amputation rates by hospital referral region, comparing unadjusted and adjusted data to determine if the variation could be explained by factors previously shown to increase risk of lower extremity amputation.
We examined trends in lower extremity revascularization procedures using models analogous to those used for examining trends in lower extremity amputation.
To estimate 1-year mortality after lower extremity amputation for each cohort year from 2000 through 2013, we adjusted for the same factors as those used in the lower extremity amputation model and performed a subanalysis by amputation level.
In modeling the primary outcome, we had 3 702 395 total observations with less than 0.1% of the data missing. Thus, we conducted a complete case analysis using 3 700 902 records. All statistical analyses were performed using SAS for Windows, version 9.4 (SAS Institute Inc) and Stata/MP, version 13.1 (StataCorp). A 2-sided P < .05 was considered statistically significant. Data analysis was conducted from August 2017 to April 2018.
In each annual cohort, there were fewer women (47.5% in 2000, 46.2% in 2005, 44.9% in 2010, and 44.0% in 2014) than men, more than half of the patients were white individuals (58.1% in 2000, 56.9% in 2005, 56.9% in 2010, and 56.7% in 2014), and a relatively small proportion were employed (13.9% in 2000, 15.1% in 2005, 16.1% in 2010, and 16.5% in 2014) (Table 1). The proportion of patients with diabetes and hypertension increased with time, whereas the proportion of patients with coronary artery disease and recognized PAD decreased with time.
The adjusted rate of all lower extremity amputations decreased from 5.42 per 100 person-years (95% CI, 5.28-5.56) in 2000 to 2.66 per 100 person-years (95% CI, 2.59-2.72) in 2014 (Figure 1; eTable 3 in the Supplement), a relative decrease of 51.0%. The adjusted rate of above-knee amputations decreased by 65.0% and the adjusted rate of below-knee amputations decreased by 58.5%, whereas the adjusted rate of below-ankle amputations decreased by only 25.9% from 2000 to 2014.
The adjusted rate of any lower extremity revascularization procedure decreased during the study period from 1.63 per 100 person-years (95% CI, 1.55-1.71) in 2000 to 1.28 per 100 person-years (95% CI, 1.24-1.33) in 2014, a decrease of 21.5% (eFigure, eTable 4 in the Supplement). The adjusted rates of surgical bypass decreased by 56.4% (95% CI, 52.5%-60.0%), whereas the adjusted rate of endovascular revascularization increased by 37.0% (95% CI, 25.5%-49.7%).
During the study period, the adjusted amputation rate for patients with diabetes decreased by 52.8% from 8.65 per 100 person-years (95% CI, 8.41-8.88) in 2000 to 4.09 per 100 person-years (95% CI, 3.99-4.19) in 2014; for patients without diabetes, the rate decreased by 48.0% from 1.43 per 100 person-years (95% CI, 1.31-1.54) to 0.74 per 100 person-years (95% CI, 0.69-0.79) (eTable 5 in the Supplement). Although the amputation rates decreased more quickly for patients with vs without diabetes (P < .001 for interaction), the amputation rates in patients with diabetes remained more than 5 times as high as those without diabetes throughout the study period (Figure 2A; eTable 5 in the Supplement). Compared with the overall cohort, patients without diabetes or hypertension had lower adjusted amputation rates, 1.03 per 100 person-years (95% CI, 0.78-1.27) in 2000, decreasing by 51.1% to 0.50 per 100 person-years (95% CI, 0.37-0.63) in 2014 (eTable 5 in the Supplement).
Adjusted amputation rates per 100 person-years were similar in 2000 for patients younger than 65 years (5.38; 95% CI, 5.18-5.57) vs those 65 years or older (5.25; 95% CI, 5.06-5.45) (Figure 2B; eTable 6 in the Supplement). However, with time, the adjusted amputation rates decreased less rapidly for patients who were younger than 65 years than for patients who were 65 years or older (eg, 2014 rates were 2.92 [95% CI, 2.84-3.00] vs 2.25 [95% CI, 2.16-2.35] per 100 person-years) (P = .005 for interaction). The adjusted amputation rates per 100 person-years for men (in 2000, 5.58 [95% CI, 5.38-5.78]; in 2014, 3.01 [95% CI, 2.92-3.10]) were higher than those for women (in 2000, 5.19 [95% CI, 4.99-5.38]; in 2014, 2.21 [95% CI, 2.13-2.30]) throughout the study period (Figure 2C; eTable 6 in the Supplement), and these rates for both groups decreased similarly with time (P = .06 for interaction).
Lower extremity amputation rates generally decreased among all hospital referral regions from 2000 through 2014, but regional variability in amputation rates persisted with time, despite adjustment for differences in patient demographic or comorbid conditions (Figure 3).
Adjusted 1-year mortality rates after lower extremity amputation for patients who receive dialysis decreased by 17% (95% CI, 14%-20%) during the study period from 52.2% (95% CI, 50.9%-53.4%) in 2000 to 43.6% (95% CI, 42.5%-44.8%) in 2013 (Table 2).
The rates of lower extremity amputations for patients with ESRD who receive dialysis decreased by 51.0% from 2000 to 2014, driven primarily by a decrease in the rate of major amputations: above-knee amputations decreased by 65.0% and below-knee amputations by 58.5% during the study period. This decrease in the amputation rate among patients with ESRD contrasts with the 69% increase in the amputation rate observed from 1991 to 199412 and mirrors the decreasing trend in amputations, which was also driven by a decrease in major amputations reported for the general Medicare population from 1996 to 2011.11
The reasons for the declining amputation rate cannot be directly gleaned from our observational analysis. We postulate that 1 factor may be more aggressive cardiovascular risk factor management given the declining prevalence of coronary artery disease and of recognized PAD observed during the study period. Improvement in foot care among patients with diabetes11 is another potential contributor; for example, in 1 large dialysis organization, implementation of monthly foot checks was associated with a 17% decrease in amputation rates.19 Although a study of patients with Medicare without ESRD showed a temporal increase in lower extremity revascularization procedures coincident with decreasing amputation rates,11 our study of patients with ESRD indicated a slight decline in rates of lower extremity revascularization during the study period.
Although the large decrease in rates of lower extremity amputation among patients with ESRD is encouraging, absolute amputation rates remained high when compared with reported rates in the general Medicare population. For example, in 2011, the amputation rate was more than 20-fold higher among patients with ESRD in our study compared with the general non-ESRD Medicare population (2.88 vs 0.12 per 100 person-years).11 Even in the relatively healthier subgroup of patients with ESRD without diabetes or hypertension, amputation rates remained markedly elevated (eg, amputation rate of 0.58 per 100 person-years in 2011). The amputation rate among patients with ESRD in our analysis (mean age, 63 years) more closely match the amputation rates reported in a very high-risk, non-ESRD patient population (ie, Medicare beneficiaries [mean age, 78 years] with recognized PAD),20 with an amputation rate of 5.79 per 100 patients in 2008. Moreover, 1-year mortality rates after lower extremity amputation, although also decreased, remained high during our study period: 52.2% in 2000 and 43.6% in 2013. Thus, although the disproportionately high amputation rate among patients with ESRD can be partially explained by the high prevalence of traditional risk factors, such as diabetes and hypertension, ESRD-specific factors, such as aggressive vascular calcification, inflammation, and uremia or even the dialysis procedure itself, may also contribute.5-10 Future studies focused on understanding the ESRD-specific mechanisms leading to lower extremity amputation are needed.
Certain patient subgroups had higher amputation rates than others. For example, the presence of diabetes conferred an amputation rate 5 times as high as that for patients without this condition, and men had higher rates of amputation than women throughout the study period. These results are consistent with previous studies showing that diabetes and male sex are significant risk factors for amputation in the non-ESRD population.8,10,21,22 We also found an increasing difference across the study period in amputation rates among patients younger than 65 years and those 65 years or older, with younger patients having higher amputation rates starting in 2005. Although older age is a known risk factor for PAD among patients without ESRD,1,22 the same may not be true for patients who receive dialysis.8,10 In a study evaluating amputation trends among patients with ESRD from 1991 through 1994, the peak amputation rate occurred between 45 and 54 years of age for patients with diabetes and between 55 and 64 years of age for patients without diabetes.12 Similarly, an international study of 29 838 patients receiving hemodialysis in the Dialysis Outcomes and Practice Patterns Study found that patients 65 to 74 years of age and those 75 years or older had significantly lower odds of amputation than patients 55 to 64 years of age.23 All of these results are consistent with reports that patients with ESRD who receive dialysis tend to experience premature vascular aging,24,25 which may contribute to more advanced PAD and hence higher rates of lower extremity amputation at a younger age. Alternatively, these results could also reflect the possibility that older patients who receive dialysis (or their physicians) may be more reluctant than younger patients to pursue amputation.
Amputation rates decreased across nearly all hospital referral regions from 2000 through 2014 but remained higher in the South and Northeast regions of the United States than in the West and Midwest regions. These regional differences persisted even after adjustment for differences in demographics, such as age, race/ethnicity, and socioeconomic status, and for differences in comorbid conditions, such as diabetes. Our results are consistent with other studies that have shown significant regional variation in amputation rates among other populations. The persistent regional variation could stem from differences not captured in our analysis, such as in intensity of vascular care,14 differences in surgical practice patterns, uncaptured socioeconomic variables, or structural differences in the health care system management of PAD or diabetes. Further studies aimed at better understanding factors contributing to regional variation are warranted.
Our study had several limitations. First, we were unable to determine laterality of amputation from billing claims, which precluded our ability to accurately determine whether a patient may have undergone multiple amputations vs multiple-stage procedures in the same calendar year. To avoid overestimating amputation rates, we counted only the highest-level amputation per year. Our results, therefore, are likely a conservative estimate of the true amputation rates. Second, we relied on the Medical Evidence Report, which generally has high specificity but low sensitivity, to ascertain comorbid conditions,26,27 thus resulting in underascertainment of comorbid conditions. This issue could potentially explain the lower prevalences of heart failure, coronary artery disease, PAD, and current smoking observed in the present study compared with those reported in other analyses.28,29 However, this issue would not be expected to substantially change our findings of persistently high amputation rates in ESRD. Finally, our analysis was limited to patients with ESRD who receive dialysis and who had Medicare Parts A and B as primary payer and thus may not be generalizable to patients who are privately insured.
We showed that patients with ESRD who receive dialysis had significant improvements in the rates of lower extremity amputation during the period from 2000 to 2014. However, amputation rates remained markedly higher than rates reported in the general Medicare population, and nearly half of all patients with ESRD who had undergone lower extremity amputation died within the first year after the procedure. Patients with ESRD and diabetes had particularly high amputation rates. We also found considerable regional variation in rates of lower extremity amputation that could not be fully explained by differences in patient demographics, socioeconomic status, or comorbid conditions. Our results highlight the need for more research on ways to prevent lower extremity amputation in this extremely high-risk population.
Accepted for Publication: April 18, 2018.
Corresponding Author: Tara I. Chang, MD, MS, Division of Nephrology, Department of Medicine, Stanford University, 777 Welch Rd, Ste DE, Palo Alto, CA 94304 (firstname.lastname@example.org).
Published Online: July 9, 2018. doi:10.1001/jamainternmed.2018.2436
Author Contributions: Dr Chang had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Concept and design: All authors.
Acquisition, analysis, or interpretation of data: Franz, Zheng, Montez-Rath, Chang.
Drafting of the manuscript: Franz, Zheng, Leeper, Montez-Rath, Chang.
Critical revision of the manuscript for important intellectual content: Chandra, Montez-Rath, Chang.
Statistical analysis: Franz, Zheng, Montez-Rath, Chang.
Obtained funding: Chang.
Supervision: Leeper, Chandra, Montez-Rath, Chang.
Conflict of Interest Disclosures: Dr Leeper reported serving as a paid consultant for Sanofi, Janssen, and Forty Seven Inc; Dr Chandra reported receiving consulting fees from Medtronic, Endologix, and Cook Medical; and Dr Chang reported receiving consulting fees from Janssen, Novo Nordisk, and Fresenius Medical Care outside the submitted work. No other disclosures were reported.
Funding/Support: This work was supported by a Mentored Patient-Oriented Research Career Development Award (K23DK095914) and a Small Research Grant (R03DK113341) from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) and by funding from the Stanford Cardiovascular Institute, all to Dr Chang.
Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.
Disclaimer: The manuscript was reviewed and approved for publication by an officer of the NIDDK. Data reported herein were supplied by the United States Renal Data System under a data use agreement (2015-50). Interpretation and reporting of these data are the responsibility of the authors and in no way should be seen as official policy or interpretation of the US government.
Create a personal account or sign in to: