2017 Volume 2 Article ID: 20170015
Objective: Studies on the outcomes of lower limb amputation have been limited by the use of selected study populations and small sample sizes. Although reamputation is an important outcome following lower limb amputation, the risk factors for reamputation remain to be elucidated. The present study was performed to identify risk factors for in-hospital death and reamputation following lower limb amputation. Methods: Using a national inpatient database in Japan, we identified 13,774 patients who underwent lower limb amputation. We examined the patients’ backgrounds and employed a multivariable logistic regression analysis to identify factors associated with in-hospital death or reamputation. Results: The average age of the 13,774 patients was 72.4 years, and 63.1% (n=8694) were male. The overall in-hospital mortality rate was 10.8% (1481/13,774). The reamputation rate was 10.1% (782/7779) for patients who initially underwent foot or transtibial amputation (18.2% [391/2148] for foot amputations and 6.9% [391/5631] for transtibial amputations). Multivariable logistic regression analysis revealed higher age, male sex, peripheral vascular disease, use of insulin, hemodialysis, and higher numbers of comorbidities as significant risk factors for reamputation or in-hospital death. Use of hemodialysis was the strongest risk factor (odds ratio, 2.10; 95% confidence interval, 1.87–2.35). Conclusions: The in-hospital mortality and reamputation rates following lower limb amputation were considerably high, reflecting the severely ill conditions of patients with advanced chronic diseases. Risk factors for in-hospital death and reamputation following lower limb amputation were identified. These should aid surgeons in determining a patient’s risk of a poor outcome and deciding on the level of amputation.
The backgrounds of patients undergoing lower limb amputation (LLA) have recently changed dramatically. From the 1970s to the 1990s, the main reason for LLA was traumatic injury.1,2,3) However, peripheral vascular disease (PVD) and diabetes mellitus (DM) have emerged as major causes of LLA. In recent studies, PVD and DM accounted for 61.0%–92.5% of LLAs.1,2,3,4)
Previous studies have also reported high postoperative mortality rates ranging from 4.2% to 13.5% among patients undergoing LLA.2,3,5,6,7,8,9,10,11,12) Reported risk factors for death after LLA include age; male sex; the site of amputation; the requirement of dialysis; and a diagnosis of PVD, DM, coronary artery disease, myocardial infarction, or renal insufficiency.2,3,5,6,7,8,9,10,11,13) However, most previous studies of LLA were limited because of the selected study populations and small sample sizes. Studies with large representative samples are required to confirm and update these previous findings.
Reamputation is another important outcome following LLA. Previous studies have reported that 18.8%–49.1% of patients undergoing LLA require reamputation. In one study, approximately half of patients requiring ipsilateral reamputation died within 3 years.14) The reported proportion of patients with DM requiring reamputation is approximately 60%.14,15,16,17,18) However, the risk factors for reamputation remain to be elucidated. The purpose of the present study was to identify risk factors for death and reamputation after LLA using a national inpatient database in Japan.
This study was approved by the Institutional Review Board of the University of Tokyo (approval number 3501). Because of the anonymous nature of the data, the need for informed consent was waived.
Data SourceWe used the Diagnosis Procedure Combination database, a nationwide administrative claims and discharge abstract database in Japan.19) All 82 university hospitals are obliged to participate in the database, whereas participation by other hospitals is voluntary. In 2012, there were approximately 1000 participating hospitals, and the number of patients included in the database was approximately 7 million, which represented approximately 50% of all admissions to acute-care hospitals in Japan.
The database includes the following data: patient’s sex and age; dates of admission and discharge; discharge status; diagnoses (main diseases, comorbidities at admission, and complications after admission); procedures coded with Japanese codes and the date on which each procedure was performed; drugs used; and the Barthel index at admission and discharge. The diagnoses were recorded as both Japanese text and International Classification of Disease, Tenth Revision, codes (ICD-10).
Patient Selection and DataWe identified patients who were admitted to participating hospitals from July 2010 to March 2014 and underwent LLA. We excluded patients undergoing toe amputation, upper limb amputation, bilateral LLA, or amputation for malignant tumors.
We examined the patients’ sex and age; diagnosis of PVD (ICD-10 codes, I70, I74, or I77); Charlson comorbidity index (CCI), which is the weighted sum of comorbid conditions and is applied for use in inpatient databases20,21); amputation level (foot [including transmetatarsal amputation, Lisfranc disarticulation, or Chopart’s disarticulation], transtibial, or transfemoral); use of insulin during hospitalization; hemodialysis during hospitalization; and discharge status (discharge to home, discharge to other facility, or death). Age was categorized as ≤40, 41–50, 51–60, 61–70, 71–80, 81–90, ≥ 91 years. The CCI was categorized as 0, 1, 2, and ≥3.
OutcomesThe outcomes of this study were in-hospital mortality and reamputation during hospitalization.
Statistical AnalysesWe employed a multivariable logistic regression analysis to identify factors associated with in-hospital death. Age, sex, diagnosis of PVD, insulin use, hemodialysis, CCI, and the level of the first amputation were entered as independent variables. We also conducted a similar analysis with death and reamputation as a composite outcome.
We considered a two-sided P value of less than 0.05 as statistically significant. All analyses were conducted using SPSS version 20.0 (IBM Corp., Armonk, NY, USA).
We identified 18,178 patients who underwent LLA. We excluded 36 patients who underwent upper limb amputation, 943 who underwent bilateral LLA, and 428 with malignant tumors. We also excluded 2997 patients with missing data on the side of the operation. Finally, we included 13,774 eligible patients in the present analyses.
The average age was 72.4 (standard deviation, 13.4) years. Table 1 summarizes the characteristics of the patients. More than 80% of the patients were >60 years old, 52.5% had PVD, 69.4% had a CCI of ≥1, 54.5% were undergoing treatment with insulin, and 28.0% were on hemodialysis. The in-hospital mortality rate was 10.8%, and an additional 49.4% of patients were transferred to another facility.
| Demographics and comorbidities | n | % | |
|---|---|---|---|
| Male | 8694 | 63.1 | |
| Age (years) | |||
| ≤40 | 259 | 1.9 | |
| 41–50 | 598 | 4.3 | |
| 51–60 | 1575 | 11.4 | |
| 61–70 | 3228 | 23.4 | |
| 71–80 | 3922 | 28.5 | |
| 81–90 | 3332 | 24.2 | |
| ≥91 | 860 | 6.2 | |
| Peripheral vascular disease | 7231 | 52.5 | |
| Insulin use | 7502 | 54.5 | |
| Hemodialysis | 3862 | 28.0 | |
| Charlson Comorbidity Index | |||
| 0 | 4216 | 30.6 | |
| 1 | 4615 | 33.5 | |
| 2 | 3320 | 24.1 | |
| 3- | 1623 | 11.8 | |
| Level of amputation | |||
| Foot | 1757 | 12.8 | |
| Transtibial | 5538 | 40.2 | |
| Transfemoral | 6479 | 47.0 | |
| Discharge status | |||
| Home | 5286 | 38.4 | |
| Facility or another hospital | 6799 | 49.4 | |
| Death | 1481 | 10.8 | |
Figure 1 shows the outcomes following amputation. Of the 2148 patients who underwent foot amputation, 320 (14.9%) and 71 (3.3%) required reamputation at the transtibial and transfemoral level, respectively. Of the 5631 patients who underwent transtibial amputation, 391 (6.9%) required reamputation. Among the 782 patients who underwent reamputation, 91 died postoperatively (11.6%). Table 2 summarizes the outcomes stratified by the amputation level at the first surgery. Overall, the composite outcome (death or reamputation) occurred in 2172 patients (15.8%).
Outcomes following amputation.
| First amputation level | n | Discharged to home, facility, or another hospital, n (%) | Reamputation, n (%) | Death, n (%) | Death or reamputation, n (%) |
|---|---|---|---|---|---|
| Foot | 2148 | 1651 (76.9) | 391 (18.2) | 135 (6.3) | 497 (23.1) |
| Transtibial | 5631 | 4799 (85.2) | 391 (6.9) | 503 (8.9) | 832 (14.8) |
| Transfemoral | 5995 | 5152 (85.9) | - | 843 (14.1) | 843 (14.1) |
| All | 13,774 | 11,602 (84.2) | 782 (5.7) | 1481 (10.8) | 2172 (15.8) |
Table 3 shows the results of the logistic regression analysis for in-hospital death. Significant risk factors included male sex, higher age, use of insulin, hemodialysis, higher CCI, and higher amputation level. Table 3 also shows the results of the logistic regression analysis for the composite outcome (in-hospital death or reamputation). Significant risk factors included male sex, higher age, PVD, use of insulin, hemodialysis, higher CCI, and first amputation level. Use of hemodialysis was the strongest risk factor for both in-hospital death and the composite outcome (odds ratios, 2.80 and 2.10, respectively).
| In-hospital death | In-hospital death or reamputation | ||||
|---|---|---|---|---|---|
| Variable | OR (95% CI) | P value | OR (95% CI) | P value | |
| Age | 1.05 (1.04–1.05) | <0.001 | 1.03 (1.03–1.04) | <0.001 | |
| Sex | |||||
| Female | Reference | Reference | |||
| Male | 1.21 (1.07–1.37) | 0.002 | 1.14 (1.03–1.26) | 0.014 | |
| Peripheral vascular disease | 1.16 (0.97–1.37) | 0.101 | 1.21 (1.05–1.39) | 0.010 | |
| Insulin use | 1.50 (1.33–1.70) | <0.001 | 1.38 (1.24–1.53) | <0.001 | |
| Hemodialysis | 2.80 (2.44–3.21) | <0.001 | 2.10 (1.87–2.35) | <0.001 | |
| Charlson Comorbidity Index | |||||
| 0 | Reference | Reference | |||
| 1 | 0.98 (0.83–1.16) | 0.848 | 0.89 (0.78–1.03) | 0.108 | |
| 2 | 1.17 (0.99–1.38) | 0.066 | 1.15 (0.99–1.33) | 0.061 | |
| 3- | 1.58 (1.31–1.90) | <0.001 | 1.50 (1.27–1.76) | <0.001 | |
| Level of amputation | |||||
| Foot | Reference | Reference | |||
| Transtibial | 1.38 (1.12–1.69) | 0.002 | 0.52 (0.45–0.59) | <0.001 | |
| Transfemoral | 2.07 (1.69–2.53) | <0.001 | 0.43 (0.37–0.49) | <0.001 | |
OR, odds ratio; CI, confidence interval.
Using a national inpatient database in Japan, we identified risk factors for death and reamputation after LLA. In previous studies, among patients undergoing LLA, the proportions that underwent hemodialysis, had DM, and had PVD were reportedly 9.0% to 26.8%, 24.6% to 92.5%, and 49.5% to 58.0%, respectively.3,6,11,22) The proportions of patients in the present large, nationally representative study that underwent hemodialysis or had DM or PVD were similar to those in the above-mentioned studies.
The reported mortality rate after LLA ranges from 6.4% to 12.4%.2,5,6,7,8,11,13,23,24,25) We also observed a high mortality rate of 10.8% in the present study. LLA is often performed as a life-saving measure for a critically ischemic or infected limb in patients with DM or PVD.1,26) These critically ill patients may not survive the perioperative period. Additionally, unresolved underlying diseases may disturb wound healing and cause recurrent infections. Furthermore, some patients may die of other complications such as cardiac events, stroke, or renal failure.5,11) These factors could account for the high mortality rate. Our multivariable analysis revealed several risk factors for death following LLA, including hemodialysis and the level of amputation. These findings are consistent with those in previous studies.2,7,8,9,11,23)
Little is known about reamputation following LLA, and the present study is the largest study of its kind. Reamputation occurred in 10.1% of patients who underwent foot or transtibial amputation during hospitalization. Although LLA is performed to control infection by sacrificing an extremity, one study revealed that infection often persists.7) In such cases, reamputation is necessary. Some small studies have suggested higher age, DM, gangrene, previous revascularization, and smoking to be risk factors for reamputation.14,18) In the present study, hemodialysis and a higher number of comorbidities were additional risk factors for the composite outcome.
The optimal level for the first amputation remains controversial. Distal amputation is desired to facilitate postoperative ambulation. However, proximal amputation is associated with a higher chance of wound healing.27) In the current study, patients whose first amputation level was the foot were more likely to require reamputation. Among patients whose first amputation level was the foot, some might have actually needed transtibial amputation and required reamputation following an optimistic prediction. Although it is sometimes difficult to predict the outcome before surgery, the risk factors identified in this study will help surgeons decide on the best level of amputation.
There are some limitations in this study. First, although a large number of acute-care hospitals were included in the database, not all patients who underwent LLA in Japan were examined. Second, some patients were excluded because of missing data on which side was operated on. Third, the precise reasons for amputation and in-hospital death could not be determined from the current data. Finally, we could not analyze the outcomes after hospital discharge because of the absence of data. However, reamputation is performed mainly during the postoperative period following the first amputation.14,15)
In conclusion, older patients and those with DM or PVD accounted for a large proportion of patients undergoing LLA, and the rates of in-hospital death and reamputation following LLA were considerably high. Higher age, male sex, use of insulin, hemodialysis, and higher numbers of comorbidities were identified as risk factors for in-hospital death or reamputation following LLA. These findings should aid surgeons in determining the risk of poor outcomes and deciding on the optimal level of amputation.
This research was supported by Grants for Research on Policy Planning and Evaluation from the Ministry of Health, Labour, and Welfare of Japan (H28-Policy-Designated-009 and H27-Policy-Strategy-011).
We have no conflicts of interest to declare.
