跳到正文
原文
Nature 健康护理主题·· 16 天前精选AI 评分75

瑞典全国下肢截肢术后死亡率观察性队列研究

Mortality and survival rates following lower limb amputation: a national observational cohort study in Sweden

AI 导读

瑞典一项全国性观察性队列研究显示,下肢截肢术后一年死亡率较高,不同截肢水平的死亡风险存在显著差异。研究发现,高截肢水平与较高的死亡风险相关,而部分足部截肢的死亡风险较低。

推荐理由

研究提供了瑞典全国范围内下肢截肢术后一年死亡率的最新数据,有助于评估截肢手术的临床效果。

正文

Introduction

Globally, an estimated 65 million people live with limb loss, and approximately 1.5 million individuals undergo amputations every year, of which 60% are lower limb amputations (LLA)1. The global incidence represents a substantial global health burden. Although incidence and trends vary considerably between regions, the three major contributing factors are:

  1. 1.

    Diabetes, which accounts for 45–70% of LLAs in many regions2.

  2. 2.

    Peripheral artery disease, including critical limb-threatening ischaemia, is responsible for most amputations among older adults3.

  3. 3.

    Trauma, which remains a dominant cause in low- and middle-income countries, including road traffic injuries, industrial accidents, and conflict-related trauma4.

In high-income countries, more than 90% of LLAs are dysvascular, primarily due to peripheral artery disease and diabetic foot complications5, whereas trauma, infections, and neoplasms account for a smaller but clinically important proportion6. Survival after LLA remains poor worldwide. A large systematic review and meta-analysis reported pooled mortality rates of 33.7%, 51.5%, 53%, 64.4%, and 80% at 1-, 2-, 3-, 5-, and 10-year follow-up, respectively7. These figures, however, largely reflect outcomes from earlier decades and may not fully reflect contemporary survival patterns owing to advances in revascularization, acute care, perioperative optimization, and rehabilitation.

More recent studies provide mixed results. For example, a study from the USA in 2023, including 26,669 people, reported a substantially lower 5-year mortality rate of 18.11%8. In contrast, another large dataset found a 5-year mortality of 51.4% among those who survived the immediate postoperative period, with chronic kidney disease and ischaemic heart disease identified as strong independent predictors of death9. These findings underscore that the leading causes of death after LLA are typically cardiovascular, and renal in nature, such as myocardial infarction, stroke, and complications of chronic heart failure, and renal disease, followed by infections. Taken together, these conflicting findings highlight significant variation between countries and emphasize the need for updated, population-based survival data.

Sweden represents a high-income country with universal health care, established vascular services, and national registers that enable comprehensive epidemiological surveillance. Approximately 1,800 initial LLAs are performed in Sweden annually (excluding toe amputations)10. These conditions make Sweden a suitable setting for nationwide assessment of contemporary mortality after LLA and for comparisons across regions.

Amputation level is a critical determinant of clinical outcomes, including postoperative mortality, rehabilitation potential, prosthetic use, functional independence, and long-term survival. Individuals undergoing LLA are predominantly older and commonly have diabetes and/or vascular disease, often combined with multiple comorbidities that negatively influence rehabilitation11. One-year mortality remains high and differs substantially by level: 29% after transtibial amputation (TTA) and 45% after transfemoral amputation (TFA), according to the latest report by SwedeAmp12, the Swedish Amputation and Prosthetics Registry. However, SwedeAmp covers less than 50% of all initial LLAs in Sweden, highlighting the need for nationwide register-based studies that include all amputation levels to accurately assess post-amputation mortality10.

Higher-level amputations, such as TFA, are linked to poorer prognosis, greater metabolic demands during ambulation, slower rehabilitation progress, and a lower likelihood of successful prosthetic use13. Conversely, preserving the knee joint through TTA is associated with markedly better survival, improved mobility outcomes, and greater independence14. In addition, variation in amputation level may reflect underlying differences in vascular disease severity, access to revascularisation, regional surgical practice, and clinical decision-making. Understanding these patterns is essential for improving treatment pathways, strengthening preventive strategies, and informing healthcare resource planning at both national and regional levels.

To address this gap and provide contemporary population-based evidence to support international benchmarking, this study examined the association between initial LLA, and mortality compared with matched controls using prospectively recorded data from Sweden between 2008 and 2017. We also assessed whether this association varied by region of residence by comparing individuals with and without LLA within each region and formally testing the interaction between LLA and region. Further, we examined temporal trends in 1-year mortality among individuals with LLA. Finally, in secondary exploratory analyses, we examined associations between selected comorbidities and mortality among individuals with LLA and described selected causes of death among individuals with and without LLA.

Materials and methods

Study design and data

A matched cohort study design was used. The study was reported in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines15. All individuals aged 18 years or older in Sweden who underwent a first LLA between 2008 and 2017 comprised the exposed group (n = 16,757 after excluding 184 individuals with an incorrect date of death). These individuals were identified using the Swedish National Patient Register (NPR), which includes records of all inpatient admissions in Sweden16. Amputation level was recorded using surgical procedure codes in the NPR.

Comparison group

For each individual with LLA, up to 10 individuals from the general population were randomly selected as a comparison group. Comparison individuals had no history of LLA and remained free of LLA during follow-up and were selected using exact matching on the following criteria:

  • same sex.

  • born in the same calendar year.

  • resided in the same county at the time of matching.

After excluding 2,265 individuals with an incorrect date of death, the comparison group consisted of 165,555 matched individuals without a history of LLA.

Data Linkage

Using the unique Swedish personal identification number assigned to all residents, the exposed and comparison cohorts were linked to the following registers:

  • The Cause of Death Register, to obtain dates and causes of death17.

  • The Longitudinal Integrated Database for Health Insurance and Labor Market Studies (LISA)18, to obtain demographic and sociodemographic factors relevant to survival analyses (marital status, education, birth in Sweden; see Supplementary Table 1).

Lower limb amputation

Amputation level was included as a key variable due to its clinical and prognostic significance. Initial amputation was identified using the Nordic Classification of Surgical Procedures codes used in the NPR and categorized into three levels (see also Supplementary Table 2):

  1. (1)

    high proximal amputations (codes NEQ19, NFQ09, NFQ19, and NGQ09).

  2. (2)

    low proximal amputations (NGQ19, NHQ09, and NHQ11).

  3. (3)

    partial foot amputations (NHQ12, NHQ13, and NHQ14).

Transpelvic amputation (NEQ19) accounted for only 57 individuals and was included within the high proximal amputation category. This classification was used to capture all register-identified LLAs in Sweden and to enable comprehensive follow-up over time. Toe amputations were not identified in the register data. Therefore, any subsequent amputation from the toe level to a more proximal level was considered the initial amputation in this study. If bilateral amputations were performed on the same date, only one amputation was included in the analysis: the most proximal amputation level, or, if both limbs were amputated at the same level, one limb was selected for consistency.

Covariables

Sweden comprises 21 self-governing healthcare regions, ranging from densely populated metropolitan regions to sparsely populated rural regions. Each is responsible for organising and providing healthcare to its population within the national publicly funded healthcare system. Age at the time of amputation, or the corresponding matching date for individuals without LLA, was categorized into 10-year intervals, and sex was categorized as female or male.

Selected comorbidities recorded as primary or secondary diagnoses in the NPR during the hospital stay that included the amputation date were identified among individuals with LLAs using the International Classification of Diseases, 10th revision codes (ICD-10). These included:

  • cancer (C00–C97).

  • diabetes (E10–E14).

  • dementia (F00–F03 and G30).

  • hypertension (I10).

  • ischaemic heart disease (I20–I25).

  • atrial fibrillation and flutter (I48).

  • heart failure (I50).

  • cerebrovascular disease (I61–I69).

  • vascular disease (I70–I89).

  • chronic obstructive pulmonary disease (J44).

  • chronic ulcer of lower limb (L97–L98).

  • chronic kidney disease (N18–N19).

Comorbidity data were not available for the comparison group and were therefore not included in the adjusted models.

This retrospective study was approved by the regional ethics review board in Uppsala, Sweden (reg. no. 2017/533 and 2017/533/1) and the need for informed consent was waived. All methods in this study were performed in accordance with the relevant guidelines and regulations.

Statistics

Descriptive statistics were used to calculate median survival and 30-day, 1-year, 3-year, and 5-year survival for people with and without LLA and by amputation level, age group, and sex. Individuals were followed from the time of LLA (or the matching date for individuals without LLA) until death, emigration, or December 31, 2018, whichever came first. Cox proportional hazards regression was used for survival analysis and to calculate hazard ratios (HRs) with 95% confidence intervals (CIs). Calculations compared people with and without LLAs; among those with amputation only, low proximal amputation was used as the reference group. Both unadjusted and adjusted models were fitted. The first adjusted model (Model 1) included age and sex, and the second model (Model 2) added sociodemographic factors (marital status, education, and birth in Sweden).

Poisson regression was used to estimate mortality rate ratios (MRRs) and 95% confidence intervals for the annual changes in one-year mortality between 2008 and 2017 for the amputation levels.

To assess whether the association between LLA and mortality differed by region, an interaction term between LLA status and region was included in the Cox regression model, and a global test of interaction was performed.

Descriptive statistics were used for exploratory analyses comparing the distribution of age and selected comorbidities between men and women and among individuals undergoing LLA and comparing the selected causes of death for individuals with LLA (in total and by amputation level) and individuals without LLA stratified by sex.

Analyses of individual comorbidities among individuals with LLA were considered exploratory. No adjustment was made for multiple comparisons; consequently, these analyses were interpreted as hypothesis-generating rather than confirmatory.

All analyses were conducted using Stata/MP 19.5.

Results

Characteristics of the study population

During the 10-year study period, 16,757 individuals underwent an initial LLA (Supplementary Table 3). The reference population consisted of 165,555 matched individuals without a history of LLA. The median follow-up was 6.0 years among individuals with LLA, and 6.0 years among individuals without LLA (reverse Kaplan-Meier method). The LLA cohort included more men (54.3%) than women (45.7%). Men were, on average, six years younger than women (75.2 vs. 81.2 years) at the time of amputation (Supplementary Table 3). Several comorbidities differed by sex; for example, vascular disease was more common among women, whereas diabetes and chronic kidney disease were more common among men.

Individuals with LLA were less often married than individuals without LLA (34.4% vs. 44.7%) and more often single (14.1% vs. 10.4%) or divorced (18.5% vs. 13.3%). They also had lower educational attainment: 52.6% of individuals with LLA had less than secondary education compared with 44.9% of individuals without LLA, while post-secondary education was less common (11.1% vs. 18.7%). The proportion born in Sweden was identical in the two groups (88.9%). Missing data were uncommon and were primarily related to education (2.0% among individuals with LLA and 1.9% among individuals without LLA), Supplementary Table 1.

Amputation and Survival

Among 16,757 individuals with LLA and 165,555 matched individuals without LLA, 12,757 (76.1%) and 62,293 (37.6%), respectively, died during follow-up (Table 1).

In the Cox regression model, LLA was strongly associated with increased mortality. In the age- and sex-adjusted model (Model 1), the hazard ratio (HR) was 4.60 (95% CI 4.51–4.69). Further adjustment for marital status, education level, and country of birth (Model 2), did not notably change the estimate (HR 4.48; 95% CI 4.39–4.57).

The mortality risk varied by amputation level, with a clear gradient across levels. In Model 1, HRs were 5.32 (95% CI 5.18–5.47) for high proximal amputation, 4.16 (95% CI 4.05–4.27) for low proximal amputation, and 4.01 (95% CI 3.77–4.27) for partial foot amputation. Corresponding estimates in Model 2 were HR 5.19 (95% CI 5.05–5.33), HR 4.05 (95% CI 3.94–4.16), and HR 3.93 (95% CI 3.69–4.19), respectively. High proximal amputation was consistently associated with the highest mortality, whereas partial foot amputation was associated with the lowest mortality risk among the LLA levels.

Amputation and survival, by sex, and age

When stratified by sex, similar associations between LLA and mortality were observed among men and women. Among men, LLA was associated with an HR of 4.68 (95% CI 4.56–4.81) after adjustment for age (Model 1) and an HR of 4.53 (95% CI 4.41–4.65) after additional sociodemographic adjustment (Model 2). Among women, the corresponding estimates were HRs of 4.51 (95% CI 4.39–4.64) and 4.43 (95% CI 4.30–4.55), respectively. In both sexes, more proximal amputations were consistently associated with higher mortality.

Stratification by age showed a clear age gradient. Among individuals aged 18–39 years, the relative mortality risk after LLA was highest, although the estimates were based on few cases and controls and had wide confidence intervals (Model 1 h 85.64; 95% CI 33.75–217.33; Model 2 h 84.77; 95% CI 32.33–222.28).

The relative mortality risk decreased progressively with increasing age. In Model 2, HRs were 38.99 (95% CI 26.01–58.43) for ages 40–49 years, 19.07 (95% CI 15.85–22.94) for ages 50–59, 10.82 (95% CI 10.07–11.62) for ages 60–69 years, 6.55 (95% CI 6.28–6.84) for ages 70–79 years, and 3.42 (95% CI 3.34–3.51) for age ≥ 80 years.

Across all age groups, high proximal LLA was associated with the highest mortality risk.

Table 1 Amputations and survival. Follow-up from amputation (corresponding date for individuals without LLA) to death, emigration, or December 31, 2018 (16,757 individuals with LLA, 165,555 without LLAs).

Full size table

Overall mortality was markedly higher among individuals with LLA than among individuals without LLA (Table 2). Median survival after LLA was 1.8 years, compared with 8.3 years among individuals without LLA. Five-year survival rate was 24.3% after any LLA and 66.2% among individuals without LLA.

Survival differed substantially by amputation level. Five-year survival was 17.3% after high proximal amputation, 27.1% after low proximal amputation, and 40.9% after partial foot amputation.

When stratified by sex, men had slightly higher survival than women. Five-year survival rate after any amputation was 27.9% in men and 20.1% in women. For high proximal amputations, five-year survival was 19.6% in men and 15.3% in women. For partial foot amputations, survival was 43.3% in men and 33.5% in women.

Short-term survival was also lower among individuals with amputation compared with individuals without LLA (Table 2). At 30 days after amputation, survival was 87.6% in the LLA group and 99.4% in the comparison group. At one year, survival was 60.0% versus 92.4%, and at three years, 37.8% versus 78.4%.

High proximal amputations were associated with the lowest short-term survival (82.5% at 30 days, 50.8% at one year, and 29.0% at three years). In contrast, partial foot amputations were associated with the highest survival (95.8%, 77.0%, and 54.7%, respectively).

Table 2 Median survival (years) and five-year survival (%).

Full size table

Over the 10-year study period, one-year mortality showed a small declining trend following high proximal amputation, with an estimated 1.4% relative reduction per year (Mortality Rate Ratio, MRR 0.986, 95% CI 0.972–0.999; p = 0.03; Fig. 1). No statistically significant temporal trends were observed for low proximal amputations (MRR 0.996, 95% CI 0.983–1.010; p = 0.60), partial foot amputations (MRR 1.009, 95% CI 0.975–1.045; p = 0.61), or all amputations combined (MRR 0.993, 95% CI 0.984–1.002; p = 0.11; not shown in the figure).

Fig. 1

One-year mortality, amputation levels, 2008–2017.

Amputation and survival in different regions

LLA was associated with substantially higher mortality compared with individuals without LLA within each of Sweden’s 21 regions (Table 3). In the fully adjusted model, regional HRs ranged from 3.67 (95% CI 3.26–4.12) in Kalmar to 5.16 (95% CI 4.56–5.83) in Södermanland. All regional estimates were above 1.0, demonstrating an increased mortality associated with LLA in every region. However, the magnitude of this association between LLA and mortality varied significantly between regions (global test of interaction, p < 0.001).

Table 3 Amputation and individuals without LLA, stratified by region. Follow-up from amputation (corresponding date for individuals without LLA) to death, emigration, or December 31, 2018 (16,757 individuals with LLA, 165,555 individuals without LLA).

Full size table

Comorbidities and survival after amputation

Exploratory analyses were performed to examine associations between age, sex, comorbidities, and mortality among individuals with LLA (Supplementary Table 4). Given the number of comparisons and the absence of adjustment for multiple testing, these results should be interpreted as hypothesis-generating.

In the multivariable model including age, sex, and comorbidities, increasing age was associated with higher mortality (HR 1.04 per year, 95% CI 1.04–1.04), whereas mortality did not differ significantly between men and women after adjustment.

Several comorbidities were associated with mortality. The strongest association was observed for chronic kidney disease (HR 1.70, 95% CI 1.61–1.79), while cancer, heart failure, dementia, and chronic obstructive pulmonary disease were also associated with higher mortality. Diabetes showed a more modest positive association with mortality, whereas hypertension was associated with lower mortality (HR 0.83, 95% CI 0.80–0.86). These secondary findings should be interpreted cautiously because of the exploratory nature of the analyses, multiple comparisons, and the potential for residual confounding.

Overall, survival after LLA remained poor. Median survival was just 1.8 years, and five-year survival was 24.3% (Table 4). Survival varied substantially by amputation level. High proximal amputations, primarily TFA, were associated with the poorest outcomes, with a median survival 1.0 years and five-year survival of 17.3%. Low proximal amputations, predominantly transtibial, were associated with moderately better survival with a median survival of 2.2 years and a five-year survival of 27.1%. Partial foot amputations were associated with the most favourable outcomes, with a median survival of 3.6 years and five-year survival rate of 40.9% (Table 4).

When median and five-year survival were examined across comorbidity groups, survival was generally higher after partial foot amputation than after more proximal amputations. This pattern was observed across several comorbidities, including diabetes, heart failure, cerebrovascular disease, and vascular disease. Individuals with dementia had particularly low observed survival across amputation levels, with five-year survival below 6% in several groups (Table 4). Given the subgroup nature of these analyses and the potential for differences in age, frailty, and other comorbidities, these findings should be considered exploratory.

Table 4 Median survival (years) and five-year survival (%), comorbidities.

Full size table

Death ascertainment

Cause-of-death information (underlying and contributing) was available for 92% of deceased individuals with LLA and 86% of deceased individuals without LLA (Supplementary Table 5). Vascular disease, diabetes, heart failure, and ischaemic heart disease were frequently recorded among individuals with LLA, whereas cancer and dementia were relatively more common among individuals without LLA. Additional analyses by sex and amputation level are presented in Supplementary Table 6. These descriptive analyses were considered exploratory.

Sex-specific analyses showed notable differences. Among men with LLA, diabetes (41.9%) and vascular disease (43.0%) were the predominant causes of death, whereas cancer, a major cause of death in the general population (26.5%), was less common among men with LLA (14.5%).

Among women with LLA, vascular disease was the leading cause of death (46.6%). Diabetes also contributed substantially (28.7%), although less frequently than among men. Dementia was less frequently recorded as a cause of death among women with amputation (13.6%) than among women without LLA (26.6%).

Diabetes-related deaths increased progressively from high proximal to partial foot amputations (27.1–53.1%), with a similar trend observed for chronic kidney disease (12.9–25.0%) (Supplementary Table 6). In contrast, deaths attributed to vascular disease ranged from 49.0% after high proximal amputation to 33.7% in partial foot amputations.

Discussion

In this nationwide matched cohort study, LLA was strongly associated with increased mortality compared with individuals without LLA. A clear gradient was observed across amputation levels, with the highest mortality following high proximal amputations and the lowest after partial foot amputations.

The excess mortality associated with LLA persisted after adjustment for measured demographic and sociodemographic factors, including marital status, education and birth in Sweden. However, comorbidity information was unavailable for the comparison group, and the relative contributions of underlying systemic disease and factors associated with the amputation itself therefore cannot be determined.

Mortality following amputation

Our findings are consistent with contemporary European registry studies showing that mortality after major LLA remains high despite improvements in vascular surgery, perioperative care, and rehabilitation. A Danish nationwide cohort study reported one-year mortality rates of approximately 20% and 46% for transtibial and transfemoral amputation, respectively19. Similarly, a long-term follow-up study from the Netherlands, including a population in which 56% had diabetes, reported a five-year mortality rate exceeding 50% among individuals with LLA20. In our cohort, 46% of individuals with LLA had diabetes, and five-year mortality after any LLA was 76%.

The high absolute mortality observed in our cohort should be interpreted in relation to both the advanced age of the study population and the duration of follow-up. During follow-up, 76.1% of individuals with LLA died; however, this proportion reflects variable follow-up according to year of cohort entry and should not be directly compared with fixed-time mortality estimates from other studies. At five years, mortality was 75.7% after LLA, compared with pooled estimates of approximately 64% after non-traumatic major LLA in a previous meta-analysis7. A more recent systematic review and meta-analysis of 140 studies of major LLA for peripheral arterial disease or diabetes, estimated mortality at 28.9% at one year and 63.0% at five years, with a median survival of 3.1 years21. The higher absolute mortality observed in our study may partly reflect the advanced age of our nationwide, unselected cohort, in which the mean age at LLA was 77.9 years. Comparisons should nevertheless be made cautiously because of differences in age, case mix, underlying disease, comorbidity, amputation level, and study design.

Survival may have improved gradually over time; nevertheless, LLA continues to represent a late manifestation of systemic atherosclerotic and metabolic diseases. The elevated mortality persisted after adjustment for measured sociodemographic factors. This indicates that the observed association was not explained by these measured factors, but the extent to which it reflects underlying cardiometabolic disease burden, disease severity, or other unmeasured clinical factors cannot be determined20,22.

Amputation level and prognosis

A pronounced mortality gradient was observed across amputation levels. Higher amputation levels are generally associated with greater tissue ischaemia, infection burden, frailty, and overall disease severity, which may partly explain their poorer prognosis. Similar gradients have been demonstrated in systematic reviews of dysvascular amputations23. Amputation level may therefore be better interpreted as a marker of underlying disease severity than as an independent causal determinant of survival22,23.

Age, sex, and sociodemographic factors

Advanced age is an important determinant of the high absolute mortality observed in both the LLA and comparison groups. The mean age of individuals undergoing LLA was 77.9 years, and the comparison group was individually matched by year of birth, sex, and county of residence. Consequently, the substantial mortality observed among individuals without LLA (37.6% during follow-up and 33.8% at five years) should be interpreted in the context of an elderly matched population with a high background mortality risk, rather than as mortality in the general Swedish population.

Importantly, however, age did not explain the markedly poorer survival associated with LLA. The overall mortality HR was 3.88 in the unadjusted model and 4.60 after adjustment for age and sex, indicating that the excess mortality associated with LLA remained pronounced after accounting for age. The age-stratified analyses further illustrate the distinction between absolute and relative mortality. Although absolute mortality increases with age, the relative mortality associated with LLA decreased progressively with increasing age. Thus, the high mortality observed in the comparison group largely reflects the advanced age of the matched population, whereas LLA identifies a group with substantial excess mortality across all age categories24.

Although men underwent amputation more frequently than women, differences in adjusted survival were small. This supports previous findings showing that women may have lower unadjusted survival after LLA, partly because they are, on average, older at the time of amputation. In our cohort, women were approximately six years older than men at the time of amputation. Comorbidity patterns also differed by sex: vascular disease and dementia were more common among women, whereas diabetes and chronic kidney disease were more common among men. Despite these differences, the association between LLA and mortality was similar in men and women after adjustment for age and sociodemographic factors. In addition, sex was not significantly associated with survival in the model restricted to individuals with LLA after adjustment for age and comorbidities.

Crude sex differences in survival may therefore partly reflect differences in age, comorbidity burden, cardiovascular risk profiles, underlying amputation aetiology, and distribution of amputation levels rather than purely biological differences25. However, these mechanisms were not specifically examined, and residual confounding by disease severity, frailty, and other unmeasured clinical factors cannot be excluded.

Adjustment for marital status, education, and birth in Sweden only modestly attenuated the association between LLA and mortality. Previous studies have reported associations between lower educational attainment, reduced social integration, and poorer outcomes after amputation26,27.

Temporal and regional variations

One-year mortality remained largely stable during the study period. There was no significant temporal trend for all amputations combined, low proximal amputations, or partial foot amputations. A small decline was observed after high proximal amputation, corresponding to an estimated 1.4% relative reduction in one-year mortality per calendar year. A similar improvement after proximal amputation was reported in the Danish study, although that study also found declining mortality after lower-level major amputations19. In contrast, no improvement was found in an Austrian single-centre study of non-traumatic LLA between 2006 and 202225. Although the decline observed in the present study may reflect improvements in patient selection, vascular and perioperative care, or management of underlying cardiovascular disease, these potential explanations were not examined in the present study and should therefore be interpreted cautiously.

Moderate regional differences in survival were observed, consistent with previous Swedish registry data12.

Sweden is divided into 21 healthcare regions that vary considerably in population size, population density, degree of urbanization (majority of people live in cities), and geographical distances to specialist care. Healthcare is predominantly publicly funded, and access is intended to be provided on equal terms irrespective of place of residence. Specialist care is concentrated mainly at larger hospitals, while patients in smaller or more rural regions can be referred to centers providing care not available locally. Multidisciplinary diabetic foot teams are generally available, although the composition may differ between hospitals; for example, vascular surgical expertise may not always be available on site at smaller hospitals.

LLA was associated with markedly higher mortality than observed among individuals without LLA in every Swedish region, although the magnitude of the association varied significantly between regions. In the fully adjusted model, HRs ranged from approximately 3.7 to 5.2. This suggests that, although the poor prognosis associated with LLA is a nationwide finding, its relative magnitude differs geographically. However, these regional differences in relative mortality should not be interpreted as differences in post-amputation survival or healthcare quality. Regional HRs depend on mortality in both the LLA and matched comparison groups, and variation in background mortality may contribute to the observed differences. Indeed, regions with higher relative HRs did not necessarily have higher observed mortality among individuals undergoing LLA. Differences in underlying disease burden, patient selection, and other unmeasured clinical characteristics may also contribute. Consequently, the present regional analysis identifies geographical heterogeneity in the prognostic association between LLA and mortality but does not establish regional differences in the effectiveness of post-amputation care.

The incidence of first-time amputations in Nordic countries is decreasing, probably due to improved revascularization and multidisciplinary diabetes care28.

Comorbidities and risk modification

Exploratory analyses indicated associations between several comorbidities and mortality among individuals with LLA. Chronic kidney disease showed a particularly strong association with mortality, consistent with previous studies of individuals with amputation and impaired kidney function29. Heart failure, cancer, dementia, and chronic obstructive pulmonary disease were also associated with poorer survival. Diabetes showed a more modest association with mortality, whereas hypertension was inversely associated with mortality. The latter finding may reflect treatment effects, residual confounding, or previously described inverse associations between blood pressure and mortality in older populations with substantial comorbidity30. Given the exploratory nature of these analyses and the multiple comparisons performed, these findings should be interpreted cautiously and require confirmation in studies specifically designed to evaluate individual prognostic factors.

Causes of death

The descriptive cause-of-death analyses showed that cardiovascular, renal, vascular and metabolic conditions, including heart failure, ischaemic heart disease, vascular disease, diabetes, and kidney disease, were frequently recorded among deceased individuals with LLA. Cancer and dementia were relatively more common among deceased individuals without LLA. These exploratory observations are consistent with a substantial cardiometabolic disease burden among individuals undergoing LLA but should not be interpreted as establishing causal relationships between amputation, comorbidity, and specific causes of death20,22,29.

Clinical implications

LLA is associated with very high mortality, and management should therefore extend beyond surgery to include secondary prevention strategies, such as aggressive cardiometabolic risk control, renal evaluation, and multidisciplinary limb care31,32.

Strengths and limitations

The major strengths of this study include its nationwide population-based design and the inclusion of all first LLAs recorded in Swedish inpatient care during the study period, which support the generalizability of the findings to adults undergoing LLA in Sweden. The large cohort, inclusion of all 21 healthcare regions, and use of national registers reduce the risk of selection related to individual hospitals or specialist centres. The findings may also be relevant to other high-income countries with similar ageing populations, universal healthcare systems, and patterns of diabetes and vascular disease. However, generalizability to countries with different healthcare systems, access to vascular and multidisciplinary care, population characteristics, or amputation practices should be interpreted cautiously.

The markedly elevated mortality associated with LLA likely reflects, at least in part, the substantial underlying systemic disease burden among individuals undergoing amputation20,22. The comparison with matched individuals without LLA should not be interpreted as estimating the causal effect of the amputation procedure. The observed association likely reflects differences in underlying disease burden and severity. Because comorbidity information was unavailable for the comparison group, these differences could not be accounted for in the main analyses. Furthermore, the available data did not permit reliable classification of the underlying indication for amputation. Consequently, we could not perform aetiology-specific comparisons between individuals with and without LLA. The findings should therefore primarily be interpreted as describing the prognostic significance of undergoing LLA rather than the independent effect of the surgical procedure.

A limitation was that the level of detail in the analyses was constrained by the fact that certain regions in Sweden have very few amputations, especially partial foot amputations.

We have not distinguished traumatic from non-traumatic lower LLAs, which is a potentially important source of heterogeneity and selection bias because the underlying mechanisms, comorbidity profiles, and prognosis of traumatic amputees may differ substantially from those of patients undergoing amputation for vascular, diabetic, renal, or other chronic disease. However, in Sweden less than 2% of all amputations are due to trauma (Swedish registry data in “SwedeAmp”)12 and, consequently in our study only 61 individuals (approximately 0.4% of the 16,757 individuals with LLA) had a trauma diagnosis associated with LLA; suggesting that traumatic amputations constituted only a small proportion of the cohort. In addition, the available data did not permit reliable distinction between acute/emergent and planned amputations or classification according to timing in relation to ischaemia or revascularization. Differences in case mix and perioperative risk between acute and planned procedures may therefore have contributed to the observed mortality. Smoking is an important potential confounder in the relationship between LLA, vascular disease, and mortality. Unfortunately, data on smoking prevalence were not available in this material. Additionally, residual confounding may have persisted despite the adjustments made.

Conclusions

Lower limb amputation was strongly associated with increased mortality compared with matched individuals without LLA, with poorer survival following more proximal amputation. The association between LLA and mortality was observed across all Swedish regions, although its magnitude varied significantly between regions. One-year mortality remained largely stable between 2008 and 2017, with a small declining trend observed following high proximal amputation. Exploratory analyses indicated that several comorbidities, particularly chronic kidney disease and heart failure, were associated with increased mortality among individuals with LLA. These findings demonstrate the substantial mortality burden following LLA and support comprehensive management addressing the underlying systemic disease burden in this population.

References

  1. Okesina, A. A., Nsubuga, I., Omoola, O. O. & Okesina, H. A. Understanding lower limb amputation: A review of the strategies for healthcare improvement, prevention, and management. Rwanda Med. J.81(1), 118–133 (2024).

    Article  Google Scholar 

  2. Humphries, M. D. Social and structural determinants of lower extremity amputations in diabetes. Curr. Diab. Rep.25(1), 40 (2025).

    Article  MathSciNet  PubMed  PubMed Central  Google Scholar 

  3. Koivunen, V. et al. Major lower limb amputations and amputees in an aging population in Southwest Finland 2007–2017. Clin. Interv. Aging. 17, 925–936 (2022).

    Article  PubMed  PubMed Central  Google Scholar 

  4. McDonald, C. L., Westcott-McCoy, S., Weaver, M. R., Haagsma, J. & Kartin, D. Global prevalence of traumatic non-fatal limb amputation. Prosthet. Orthot. Int.45(2), 105–114 (2021).

    Article  PubMed  Google Scholar 

  5. Alaseem, A. et al. Etiologies and trends in extremity amputations: A 10-year single-center experience. Healthc. (Basel Switzerland). 13, 18 (2025).

    Google Scholar 

  6. Lazzarini, P. A., O’Rourke, S. R., Russell, A. W., Clark, D. & Kuys, S. S. What are the key conditions associated with lower limb amputations in a major Australian teaching hospital? J. Foot Ankle Res.5(1), 12 (2012).

    Article  PubMed  PubMed Central  Google Scholar 

  7. Meshkin, D. H. et al. Long-term mortality after nontraumatic major lower extremity amputation: A systematic review and meta-analysis. J. Foot Ankle Surg. Offic. Publ. Am. Coll. Foot Ankle Surg.60(3), 567–576 (2021).

    Google Scholar 

  8. Beeson, S. A. et al. Analysis of 5-year mortality following lower extremity amputation due to vascular disease. Plast. Reconstr. Surg. Global open.11(1), e4727 (2023).

    Article  Google Scholar 

  9. Pande, S. D., Kamal, A., Zaw, E. & Tin, A. S. Patients following lower limb amputation: A retrospective cohort study showing how to improve survival and rehabilitation outcomes. Int. J. Surg. Open.20, 24–28 (2019).

    Article  Google Scholar 

  10. Johannesson, A. G., Scheving, R., Westlund, K. L. & Fridriksson, T. Evaluation of the SwedeAmp database: Focus on coverage and amputation level rates. Can. Prosthetics Orthot. J.7(2), 44089 (2024).

    Article  Google Scholar 

  11. Jarl, G., Johannesson, G. A., Carlberg, M., Jansson, S. P. O. & Hiyoshi, A. Editor’s Choice—Incidence of lower limb amputations in Sweden from 2008 to 2017. Eur. J. vascular Endovasc. Surg. Off. J. Eur. Soc. Vasc. Surg..64 (2–3), 266–273 (2022).

    Article  Google Scholar 

  12. SwedeAmp Annual report. The Amputation and Prosthetics Registry for the lower Extremity Lund2026 [updated September 10th 2026. Access date September 17, 2026; cited 2026 September 17]. (2025). Available from: https://rcsyd.se/swedeamp/wp-content/uploads/sites/21/2026/09/Arsrapport-SwedeAmp-2025.pdf

  13. Miller, T. A., Paul, R., Forthofer, M. & Wurdeman, S. R. Impact of time to receipt of prosthesis on total healthcare costs 12 months postamputation. Am. J. Phys. Med. Rehabil.99(11), 1026–1031 (2020).

    Article  PubMed  PubMed Central  Google Scholar 

  14. Miller, T. A., Paul, R., Forthofer, M. & Wurdeman, S. R. Factors that influence time to prosthesis receipt after lower limb amputation: A Cox proportional hazard model regression. PM R: J. Injury Funct. Rehabil.. 15(4), 474–481 (2023).

    Article  Google Scholar 

  15. von Elm, E. et al. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: Guidelines for reporting observational studies. J. Clin. Epidemiol.61(4), 344–349 (2008).

    Article  Google Scholar 

  16. Everhov, Å. H. et al. Diagnostic accuracy in the Swedish national patient register: A review including diagnoses in the outpatient register. Eur. J. Epidemiol.40(3), 359–369 (2025).

    Article  PubMed  PubMed Central  Google Scholar 

  17. Brooke, H. L. et al. The Swedish cause of death register. Eur. J. Epidemiol.32(9), 765–773 (2017).

    Article  PubMed  PubMed Central  Google Scholar 

  18. Ludvigsson, J. F., Svedberg, P., Olén, O., Bruze, G. & Neovius, M. The longitudinal integrated database for health insurance and labour market studies (LISA) and its use in medical research. Eur. J. Epidemiol.34(4), 423–437 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  19. Brix, A. T. H., Rubin, K. H., Nymark, T., Schmal, H. & Lindberg-Larsen, M. Mortality after major lower extremity amputation and association with index level: A cohort study based on 11,205 first-time amputations from nationwide Danish databases. Acta Orthop.95, 358–363 (2024).

    Article  PubMed  PubMed Central  Google Scholar 

  20. Rosien, L. et al. Long-term mortality rates after lower extremity amputation in individuals with and without diabetes mellitus (DUDE-10): A retrospective matched cohort study. J. Diabetes Complic.. 39(3), 108956 (2025).

    Article  Google Scholar 

  21. Leatherby, R. J. et al. Editor’s Choice—systematic review, meta-analysis, and time to event analysis of contemporary mortality after major lower limb amputation for peripheral arterial disease or diabetes mellitus. Eur. J. Vasc. Endovasc. Surg. Off. J. Eur. Soc. Vasc. Surg.71(5), 781–792 (2026).

    Article  Google Scholar 

  22. Calabrese, L., Maffoni, M., Torlaschi, V. & Pierobon, A. What is hidden behind amputation? quanti-qualitative systematic review on psychological adjustment and quality of life in lower limb amputees for non-traumatic reasons. Healthc. (Basel Switzerland)11(11), 1661 (2023).

  23. Ward, Z., Ridgewell, E., Quigley, M., Fatone, S. & Dillon, M. P. Proportionate mortality following dysvascular partial foot amputation and how this compares to transtibial amputation: A systematic review. Disabil. Rehabil.47(3), 549–559 (2025).

    Article  PubMed  Google Scholar 

  24. Fard, B., Dijkstra, P. U., Voesten, H. & Geertzen, J. H. B. Mortality, Reamputation, and preoperative comorbidities in patients undergoing dysvascular lower limb amputation. Ann. Vasc. Surg.64, 228–238 (2020).

    Article  PubMed  Google Scholar 

  25. Kaser, S. et al. Non-traumatic lower-limb amputations: Outcome, sex-differences, comorbidity patterns and temporal trends from 2006 to 2022. J. Clin. Med.14(12), 4030 (2025).

    Article  PubMed  PubMed Central  Google Scholar 

  26. Chernev, I. & Chernev, A. Education level among patients with major limb amputation. Cureus12(4), e7673 (2020).

    PubMed  PubMed Central  Google Scholar 

  27. Hawkins, A. T. et al. The effect of social integration on outcomes after major lower extremity amputation. J. Vasc Surg.63(1), 154–162 (2016).

    Article  PubMed  Google Scholar 

  28. Brix, A. T. H., Rubin, K. H., Nymark, T., Schmal, H. & Lindberg-Larsen, M. Epidemiology of first-time major lower extremity amputations-A Danish Nationwide cohort study from 2010 to 2021. Eur. J. Epidemiol.40(3), 297–307 (2025).

    Article  PubMed  PubMed Central  Google Scholar 

  29. Pitsenbarger, L. T. et al. Moderate and severe chronic kidney disease predict greater 5-year mortality following major lower-extremity amputation. Ann. Vasc. Surg.105, 307–315 (2024).

    Article  PubMed  Google Scholar 

  30. Chobanian, A. V. Shattuck Lecture. The hypertension paradox–more uncontrolled disease despite improved therapy. N Engl. J. Med.361(9), 878–887 (2009).

    Article  PubMed  Google Scholar 

  31. Conte, M. S. et al. Global vascular guidelines on the management of chronic limb-threatening ischemia. J. Vasc. Surg.69(6), 3S–125Se40 (2019).

    Article  PubMed  PubMed Central  Google Scholar 

  32. Kawaji, Q. et al. Multidisciplinary limb salvage care is associated with decreased mortality without increasing revascularization in major amputations. J. Vasc Surg.82(4), 1438–47e1 (2025).

    Article  PubMed  Google Scholar 

Download references

来源:Nature 健康护理主题 · nature.com