Declining Amputation Rates Secondary to Diabetes and PVD: The Role of Time, Geography, Risk, and Race

Content provided by The O&P EDGE
Current Issue - Free Subscription - Free eNewsletter - Advertise

While it is generally understood within the healthcare environment that the prevalence of diabetes and peripheral vascular disease (PVD) are on the rise, the correlating assumption that this is leading to ever-increasing numbers of lower-limb amputations does not appear to hold true. Aggressive screening and improved management strategies for patients with diabetes and PVD have actually resulted in a decline in lower-limb amputations within these populations. However, the rate of this decline appears to be variable. This article reviews the temporal trends in amputation rates in patients with diabetes along with the effects of geographic variation, associated comorbidities, and race.


While the prevalence of diabetes continues to increase in the United States, recent data suggests that the amputation rate in this population has declined sharply over the past 20 years. In a representative study, researchers recently examined healthcare data from 1988 to 2008.1 The rate of diabetes was determined from the National Health Interview Survey (NHIS), a survey of civilian, noninstitutionalized households that collects health information regarding the prevalence and incidence of disease, the extent of disability, and the utilization of healthcare services. As might be expected, the number of U.S. residents diagnosed with diabetes increased substantially over the 20-year period, from 5.4 million in 1988 to 17.1 million in 2008.1

Nontraumatic amputation rates during that same time period were extrapolated through an examination of the National Hospital Discharge Survey (NHDS). This probability survey, drawn from hospitals in all 50 states, collects patient discharge data including age, race, diagnoses, and the surgical procedures they underwent. Using International Classification of Diseases, Ninth Revision (ICD-9) procedure codes, the study's authors were able to track the number of nontraumatic lower-limb amputations that were performed over the 20-year period. Using ICD-9 diagnosis codes, these could be further refined with respect to the presence or absence of comorbid diabetes.1 Dividing the number of diabetes-related, nontraumatic lower-limb amputations from the NHDS by the relative incidence of diabetes as drawn from the NHIS allowed for calculations of the relative rate of amputations in this population across the 20-year period.1

Using this methodology, authors were able to track a steady increase in the amputation rate in the early 1990s followed by an 8.6 percent annual decline in this rate from 1996 to 2008.1 Taken cumulatively, the authors cited a 67 percent decline in the nontraumatic lower-limb amputation rate during this period.1

These findings are not unique in recent literature. In their conclusions, the authors of the present study cited similar findings in recent analyses of people with diabetes in general, in the Medicare population, in patients served by the U.S. Veterans Health Administration, and in studies drawn from Finland and England.1 The amputation rate secondary to diabetes has been in steady decline for more than a decade.


A similar analysis was recently performed on the decreasing rate of lower-limb amputations secondary to peripheral artery disease (PAD).2 In this case, the data was drawn from 2000 to 2008 Medicare files using similar ICD-9-based methodologies. Comparable to the previous study, the authors observed that from 2000 to 2008 the rate of lower-limb amputations per 100,000 Medicare beneficiaries with PAD decreased by about 20 percent, from 7,258 to 5,790.2

Additional insights were gained as the authors subdivided their findings by clustering the data according to the nine geographic divisions used by the U.S. Census Bureau, observing that amputation rates varied widely by region2 (Figure 1).

Figure 1: Geographic and temporal amputation rates among Medicare beneficiaries with PAD; number of lower-limb amputations per 100,000. Data extracted from Jones et al., 2012.2

REGION 2000 2004 2008
Highest Rates
†††West South Central 9.237 8.256 6.874
†††East South Central 8.268 8.797 7.591
†††South Atlantic 7.959 7.779 6.198
Lowest Rates
†††West North Central 5.934 5.606 4.804
†††Mountain 6.134 5.182 5.052
†††New England 6.484 5.267 5.006
†††East North Central 6.511 5.791 5.043

While the amputation rate has been dropping across the United States, it has remained comparatively higher in the West South Central (Arkansas, Louisiana, Oklahoma, and Texas); East South Central (Alabama, Kentucky, Mississippi, and Tennessee); and South Atlantic (Delaware, District of Columbia, Florida, Georgia, Maryland, North Carolina, South Carolina, Virginia, and West Virginia) regions. Recalling that this data represents the rate of amputations due to PAD rather than the incidence of PAD itself, it suggests that individuals with PAD in these regions undergo amputations more frequently than individuals with PAD in other regions.

By contrast, the lowest rates were found in West North Central (Iowa, Kansas, Minnesota, Missouri, Nebraska, North Dakota, and South Dakota). In the year 2000, this was closely followed by the Mountain region (Arizona, Colorado, Idaho, Montana, Nevada, New Mexico, Utah, and Wyoming). Because of sharper regional declines, by 2008 lower amputation rates were also observed in New England (Connecticut, Maine, Massachusetts, New Hampshire, Rhode Island, and Vermont) and East North Central (Illinois, Indiana, Michigan, Ohio, and Wisconsin).2

The authors cited potential explanations for this regional variation, several of which will be discussed later in greater detail. These include variations in early screening and detection, revascularization procedures, patient education, timing of initial presentation for evaluation, and thresholds to perform lower-limb amputation due to PAD.2


A similar locational analysis was performed with respect to the amputation rate among patients with diabetes.3 Confined to a tighter temporal window of Medicare beneficiaries undergoing lower-limb amputations from 2006 to 2008, the authors chose to perform their analysis according to the 306 geographic divisions of the Dartmouth Atlas of Health Care, referred to as Hospital Referral Regions (HRRs).3

Even when limiting their examination to three years, the authors confirmed a declining rate of amputations during that timeframe. The choice to use smaller geographic units permitted the identification of smaller pockets throughout the country where HRRs had higher or lower amputation rates. For example, pockets with the highest amputation rates were observed in southern Texas, southwestern Arkansas, southeastern Louisiana, east central Mississippi, and southwestern Georgia. Amputation rates in these regions were three to four times higher than those observed in places like southern Nevada, northern Montana, eastern South Dakota, northern Michigan, or southern Florida.3

Unique to this study was the observation that lower-limb amputation rates secondary to diabetes were more similar in neighboring HRRs than those HRRs that were farther apart.3 That is, HRRs with higher amputation rates were generally near HRRs with similarly higher rates, and HRRs with lower amputation rates were generally near HRRs with similarly lower rates. Using this analysis, disproportionately high amputation rates among Medicare beneficiaries with diabetes were observed in a continuous region containing southeast Texas, southern Oklahoma, Louisiana, Arkansas, and Mississippi. By contrast, disproportionately lower amputation rates were observed in continuous HRR clusters located in portions of southern Florida, parts of New Mexico and Arizona, and Eastern Michigan.3

Possible explanations for this geographic variation included localized practice patterns, thresholds for performing amputations, localized trends in training standards, regional variations in socioeconomic status, patient-based cultural or educational differences, and variability in healthcare access and preventive care.3


As indicated in the beginning of this article, geographic variations aside, amputation rates secondary to diabetes and PVD are on the decline. The next article under consideration sought to determine whether comorbidities played a role in the likelihood of lower-limb amputations within the population of persons with diabetes.4 While their sampling approach differed slightly from those described so far, the authors of this study also drew from Medicare data sets, in this case collecting information on individuals who underwent lower-limb amputations between 1999 and 2006.

The authors collected data on basic demographics such as age, sex, and race. In addition, they analyzed the cohort with respect to the presence of comorbid stroke, congestive heart failure, myocardial infarction, ischemic heart disease, and chronic kidney disease. This allowed them to separate individuals into one of two cohorts. The "high-risk" cohort was defined as those individuals with diabetes and either two additional comorbidities, end-stage renal disease (ESRD), or both. The "low-risk" cohort had fewer comorbidities and did not have ESRD.4

As with the previous studies, a progressive decline in the rate of amputations among patients with diabetes was observed. Additionally, the authors observed a slight decrease between 1999 and 2006 in the average age of individuals who underwent amputations. This was fueled by an increase in the number of amputations occurring in individuals between the ages of 50 and 64 (16 percent to 22 percent) and a slight drop in individuals between the ages of 65 and 74 (35 percent to 31 percent).4 Further, there was an increase in the proportion of amputations among the high-risk cohort over time. Despite representing a minority of the overall population of individuals with diabetes (4 percent in 1999 and 10 percent in 2006), this cohort experienced one-third of the lower-limb amputations in 1999 and half of the amputations in 2006. Viewed another way, the amputation rate among the high-risk patients (28.6 per 1,000) was more than nine times higher than that observed among the lowrisk patients (3 per 1,000).4 For the clinical prosthetist, these observations suggest a change in demographics in which individuals with vascular-related amputations may tend to be younger but present with more complex medical histories.


In addition to the changes observed over time, with respect to geography and in the presence of cumulative comorbidity, race has consistently stood out as an influential factor in amputation rates among patients with diabetes and PVD. Li et al. observed that while the amputation rates among white and black subpopulations in 1988 were similar (7.8 and 8.4 per 1,000 persons with diabetes respectively), the rate of amputation decreased more drastically within the white demographic such that by 2008 the incidence of amputation among black patients was nearly double that observed in white patients (4.9 and 2.9 per 1,000 persons with diabetes respectively).1 In their analysis, Jones et al. observed that while black patients comprised roughly 11 percent of the nearly three million patients with PVD, they accounted for just under 30 percent of the lower-limb amputations performed.2 Further, the amputation rate among Goldberg et al.'s high-risk black population was double that observed in high-risk white patients (48.9 and 24.4 per 1,000 respectively).4 Similar trends were observed in the low-risk cohorts where the amputation rate among the black demographic was 5.7 per 1,000 compared to only 2.6 per 1,000 in the white demographic.4

Possible explanations for these trends are largely speculative and include socioeconomic factors, community views, and access to preventive medicine and healthcare. A recent study by Holman et al. provides further insights by observing that black individuals were less likely than their white peers to have undergone revascularization procedures or had already experienced previous limb-related hospital admissions, toe amputations, or wound debridement prior to their major limb amputations.5 Succinctly put and generally speaking, black patients more frequently undergo amputation and white patients more frequently undergo revascularization.


Collectively, the data from these studies suggests that the lower-limb amputation rate secondary to diabetes and PVD has shown a steady decline in recent years. However, the rate of this decline appears to vary from region to region, with additional variability appearing as the product of comorbid risk factors and race. These trends will affect the future demand for prosthetic services across the country as well as the demographic characteristics that are likely to be encountered with greater regularity.

Phil Stevens, MEd, CPO, FAAOP, is in clinical practice with Hanger Clinic, Salt Lake City, Utah. He can be reached at


  1. Li, Y., N. R. Burrows, E. W. Gregg, A. Albright, and L. S. Geiss. 2012. Declining rates of hospitalization for nontraumatic lower-extremity amputation in the diabetic population aged 40 year or older: U.S., 1988-2008. Diabetes Care 35 (2):273-7.
  2. Jones, W. S, M. R. Patel, D. Dai, S. Subherwal, J. Stafford, S. Calhoun, and E. Peterson. 2012. Temporal trends and geographic variation of lower-extremity amputation in patients with peripheral artery disease: Results from U.S. Medicare 2000-2008. Journal of the American College of Cardiology 60 (21):2230-6.
  3. Margolis, D. J., O. Hoffstad, J. Nafash, C. E. Leonard, C. P. Freeman, S. Hennessy, and D. J. Wiebe. 2011. Location, location, location: Geographic clustering of lower-extremity amputation among Medicare beneficiaries with diabetes. Diabetes Care 34 (11):2363-7.
  4. Goldberg, J. B., P. P. Goodney, J. L. Cronenwett, and F. Baker. 2012. The effect of risk and race on lower extremity amputations among Medicare diabetic patients. Journal of Vascular Surgery 56 (6):1663-8.
  5. Holman, K. H., P. K. Henke, J. B. Dimick, and J. D. Birkmeyer. 2011. Racial disparities in the use of revascularization before leg amputation in Medicare patients. Journal of Vascular Surgery 54 (2):420-6.

Bookmark and Share