Factors Related to Reamputation in Patients with Diabetes Who Have Undergone Minor Amputation: A Single-Facility Retrospective Cohort Study

Background: With aging, there has been an increase in the number of patients with lower limb amputations, especially in Japan. Previous studies have reported that minor amputation after limb salvage is associated with a high recurrence, with the ulcer recurring within 10.5 months. Age, wounded area, history of peripheral artery disease, and wound infection have been reported as the risk factors of recurrence. Thus far, there are no studies on the association between rehabilitation and reamputation. This study therefore aimed to examine factors predictive of reamputation in patients with diabetes who have undergone minor amputation for developing a rehabilitation program. Methods: This study included 129 patients who had undergone physical therapy between January 2015 and February 2018. The patients were assigned to the reamputation and non-reamputation groups. We retrospectively studied the patients’ medical records and compared the factors between the groups. Next, we analyzed the differences between the factors using Cox proportional hazards analysis. Results: During the follow-up period, 42 patients (35.2%) underwent reamputation. The factors associated with reamputation were hemodialysis, ankle dorsiexion angle, and ambulation Functional Independence Measure score. Conclusions: We identied hemodialysis, ankle dorsiexion angle, and FIM ambulation were the factors associated with reamputation in patients with diabetes who had undergone minor amputation (cid:0) Future developing a rehabilitation program focusing on these factors could reduce reamputations.


Background
In Japan, the most frequent cause of amputation is trauma. However, the percentage of amputations due to peripheral circulatory disorders such as arteriosclerosis obliterans and diabetic gangrene has recently increased [1].
In a large-scale epidemiological investigation of the incidence of lower extremity amputations in diabetic patients, the annual incidence was estimated to be 0.05% [2]. In addition, a survey conducted by The Global Burden of Disease study reported that the number of patients with diabetes-related lower-extremity amputation was estimated to be 50,000 in Japan [3].
As Japan is a developed country with a growing elderly population, an increased number of patients with lower limb amputations may result in a greater burden on social welfare and medical expenses. Progress in multidisciplinary treatment has resulted in limb salvage and reduced rates of major amputation in patients with diabetes and peripheral arterial disease [4]. However, patients who had undergone minor amputation after limb salvage treatment presented with a high recurrence rate, and the ulcer has been reported to develop in 69.5% of such cases within 10.5 months [5]. It was reported that 49.1% of such patients underwent ipsilateral reamputation within 3 years after their rst surgery, with 78.9% of these undergoing reamputation in the rst 6 months [6].
Previous studies have indicated that the risk factors for recurrence in minor amputation cases are age [7], wounded area [8], history of peripheral artery disease [9], and wound infection [10].
The multidisciplinary approach for diabetic foot treatment includes debridement of the wound, management of any infection, revascularization procedures, and postoperative lower limb rest and o oading [11].
Rehabilitation is increasingly performed after amputation to prevent complications and maintain limb function.
If the association between rehabilitation and reamputation was reported, effective interventions could be applied to decrease reamputation. However, there have been no such reports to date. Therefore, this study aimed to examine the predictive factors related to minor reamputation in patients with diabetes to develop a rehabilitation program that would consider the risk of reamputation.

Study participants
This retrospective cohort study included 245 patients with diabetic ulcers who had undergone a minor amputation followed by physical therapy between January 2015 and February 2018. All subjects were patients admitted at the Oita Oka Hospital. We excluded patients with (1) infections after minor amputation, (2) major amputations (below and above the knee), (3) death or discharge due to systemic complications, (4) use of a wheelchair for mobility before hospitalization, (5) severe progression of dementia, (6) missing data, (7) and no follow-up survey data. Finally, 129 patients were enrolled in this study ( Fig. 1).
Major and minor amputations were de ned as those proximal and distal, respectively, to the ankle joints, including the toes. Reamputation was de ned as occurring in the ipsilateral limb within 1 year after the rst amputation.

Data Collection
Measurement items included participants' basic and medical information, physical function, and the presence or absence of reamputation. Basic information included age, sex, body mass index (BMI), hospitalization, physical therapy period, average length of daily physical therapy in minutes, foot o oading period, comorbidities (hypertension, hyperlipidemia, and heart disease, cerebrovascular disease, chronic kidney disease, and chronic obstructive pulmonary disease), and presence or absence of hemodialysis. The medical information included laboratory parameters (serum albumin, serum hemoglobin, blood glucose, C-reactive protein, white blood cell counts, and estimated glomerular ltration rate), lower limb blood ow data (skin perfusion pressure and ankle-brachial pressure index), wound ischemia foot infection (WIfI) classi cation system, amputation region (toe, ray, and transmetatarsal), and foot deformation (Charcot joint, hallux valgus, hammer toe, and claw toe). The physical function included knee extension muscle strength, ankle dorsi exion angle, ankle plantar exion angle, plantar sensory disorder, and Functional Independence Measure (FIM) ambulation. BMI was calculated by dividing the body weight (kg) by the square of the height (m) and was reported in kg/m 2 . The o oading period started from the minor amputation to the start of loading.
The average length of daily physical therapy was calculated by dividing total hours of physical therapy by the number of days in the hospital. The estimated glomerular ltration rate (eGFR) was calculated based on the new Japanese coe cient Modi cation of Diet in Renal Disease study equation: eGFR (mL/min/1.73 m 2 ) = 194 × (serum creatinine) − 1.094 × (age) − 0.287 (× 0.739 for females) [12].
We considered angina pectoris, myocardial infarction, percutaneous coronary intervention, or coronary artery bypass grafting as cardiovascular condition-related history and transient ischemic attack, cerebral infarction, or cerebral hemorrhage as cerebrovascular condition-related history. Lower limb blood ow data with skin perfusion pressure as a measure of blood ow were evaluated using a Laser Doppler (SensiLase PAD4000, KanekaMedix, Osaka, Japan). The ankle-brachial index was measured with an automated oscillometric device provided by Omron Colin Co., Ltd. (Tokyo, Japan). We evaluated the wound data using the WIfI and the ischemia and foot infection data using three factors. Severity was classi ed by the depth, location, and presence or absence of necrotic tissue [14]. Amputation region and foot deformation were evaluated using simple radiography and computed tomography images before and after surgery.
For physical function evaluation, we examined the maximum voluntary isometric knee extension muscle strength using a hand-held dynamometer (µ-tasF-1, Anima, Tokyo, Japan). For knee extensor strength measurements, participants were asked to sit on a chair with the knee exed at 90 degrees and push at maximum strength against the dynamometer pad for 5 s. Isometric knee extensor strength was measured twice per side, and the highest value for the right and left legs was used to represent the knee extensor muscle strength.
The range of motion in the ankle joint was evaluated by measuring dorsi exion and plantar exion on the wound side with a goniometer and the angle that allowed maximum pain-free movement. The presence or absence of plantar sensory impairment was considered as neuropathy when the evaluator size of the Semmes-Weinstein-mono lament was 5.07.
Based on previous studies, we performed three-site tests involving the plantar aspects of the great toe, third metatarsal, and fth metatarsal [15]. The ambulation status was evaluated using the movement parameter of the FIM score [16].

Intervention
In this study, the goal of rehabilitation intervention was to improve deconditioning and walking ability and expand the activities of daily living during hospitalization. The 1st postoperative day started with strength training and range-of-motion exercises of the hip and knee joints, which were performed according to the pain indication. Additionally, a walking practice started after wound healing.

Statistical analysis
Characteristic data were compared using the Mann-Whitney U test, t-test, and χ 2 test, as appropriate. A univariate Cox proportional hazards analysis was used to calculate the reamputation hazard ratio (HR) and its 95% con dence interval. A multivariate Cox regression analysis was used to determine the factors associated with reamputation. We selected the following factors for univariate analysis: age, sex, serum albumin, knee extension muscle strength, hemodialysis, ankle dorsi exion angle, and FIM ambulation.
Subsequently, a multivariate Cox regression analysis with stepwise selection was carried out to identify signi cant independent factors.
The incidence of reamputation presence and absence was calculated using the Kaplan-Meier curves for the extracted factors. The intergroup differences were estimated using the log-rank test. All analyses were performed with R version 3.2.5 (R Foundation for Statistical Computing, Vienna, Austria). The level of signi cance was set at P < 0.05.

Results
The demographic and medical information of the patients is shown in Table 1. There were 42 patients in the reamputation group within 1 year of discharge and 87 in the non-reamputation group. The reamputation group exhibited signi cantly higher rates of hemodialysis and FIM ambulation and lower rates of lower ankle dorsi exion angle than the non-reamputation group (Table 1). Univariate Cox regression analysis showed that knee extension muscle strength, hemodialysis, ankle dorsi exion angle, and the FIM ambulation administered were potential risk factors for reamputation. Subsequent multivariate Cox regression analysis with stepwise selection identi ed hemodialysis (HR 2.20, 95% CI 1.12-4.34), ankle dorsi exion angle (HR 5.82, 95% CI 2.93-11.58), and FIM ambulation (HR 3.85, 95% CI 2.00-7.39) as signi cant risk factors for reamputation ( Table 2). Figure 2, 3, and 4 shows the Kaplan-Meier curves indicating the cumulative incidence of reamputation after minor amputation. In the survival analyses using the Kaplan-Meier log-rank test, presence or absence of hemodialysis (Fig. 2), ankle dorsi exion angle (Fig. 3), and FIM ambulation (Fig. 4) were signi cant (P < 0.05).

Discussion
The present study examined factors that in uence reamputation within 1 year of discharge in patients who had undergone minor amputation. We revealed that the presence or absence of hemodialysis, ankle dorsi exion angle, and FIM ambulation were associated with reamputation in this patient population. The recurrence rate was 34.5% within 1 year of discharge, similar to previous ndings [17].
Comparing the amputation region, toe amputation showed a high reamputation rate. Regarding the relationship between amputation region and reamputation, a previous study reported that a second amputation was required in 60% of great toe amputation cases after 10 months [18]. Moreover, adjacent toe deformation has been reported as a factor in uencing reamputation after toe amputation [19]. In this study, all patients underwent physical therapy, and, therefore, the reamputation rate could be suppressed by preventing deformation. Additionally, high reamputation rates have been reported after transmetatarsal amputation in contrast to our results [20].
In recent years, the wound healing rate after transmetatarsal amputation has improved due to the use of advanced treatment techniques, and even if an increase in the number of surgeries is expected, walking ability can be maintained compared to major amputation [21]. Several previous studies have reported that limited ankle joint mobility may increase plantar mechanical stress [22,23], especially limited ankledorsi exion. The range of motion has been reported to increase the maximum plantar pressure during walking and cause pressure concentration on the forefoot [24,25]. In this study, the limited ankle joint mobility caused pressure concentration on the foot during walking, which could have resulted in wound worsening after discharge.
Regarding the relationship between hemodialysis and reamputation, it has been reported that hemodialysis caused periodic uid uctuations and worsened microcirculation, thereby promoting blood circulation disorders [26]. Miyajima et al. reported that hemodialysis was also an independent risk factor of major limb amputation [27]. Okamoto et al. reported that approximately 40% of 140 patients who had undergone hemodialysis had a peripheral arterial disease [28].
These ndings were similar to ours'. We, therefore, determined that patients who had undergone hemodialysis may have had been wounded and may have undergone reamputation due to peripheral arterial disease. Previous studies have reported the effects of different walking styles and speeds on plantar mechanical stress [29]. In addition, a study reported that an increase in cumulative plantar tissue stress associated with the extension of walking distance resulted in a wound [30].
In this study, the ambulation status was evaluated using the FIM score. However, it was di cult to compare our results with those of previous studies using the index of walking speed and style. Patients with FIM score < 5 were able to walk 50 m independently and were expected to engage in intensive walking after discharge. We also found that hemodialysis, ankle dorsi exion angle, and FIM ambulation at the time of discharge were related to reamputation rates within 1 year.
Finally, we highlight that the prevention of recurrence has become an important outcome with the progress of limb salvage treatment. This nding suggests that early follow-up is necessary for foot care and rehabilitation in patients who undergo minor amputation and engage in intensive walking.
However, this study has several limitations. First, our results were obtained from a single facility. Similar ndings derived from other facilities are needed to validate ours for generalizability. Second, the de nition of reamputation was restricted to only the original hospital. Further, patients who had not received physical therapy were excluded. Besides, we did not investigate the living conditions and selfmanagement status of patients after discharge from the hospital.
Moreover, the results may not be generalizable to all hospitalized patients with minor amputation; thus, further longitudinal studies with larger samples in a multiple hospital setting are required to investigate the reamputation rates in hospitalized patients with minor amputation.

Conclusions
To conclude, hemodialysis, ankle dorsi exion angle, and FIM ambulation were the factors associated with reamputation in patients with diabetes who had undergone minor amputation. The data obtained from this study will be valuable in aiding future rehabilitation program of preventing reamputation. Further studies with larger samples are needed to con rm our results. drafting the manuscript. All authors have read and approved the nal manuscript.

Ethics approval and consent to participate
This study was conducted in accordance with the Declaration of Helsinki (as revised in Brazil in 2013), and approval was obtained from the Ethical Committee of the Oita Oka Hospital (approval number: B0018). In place of informed consent from every participant, consent was sought by publishing details of the study on the homepage of the research institute website, as well as on-site posting in the facility Consent for publication Not applicable.    Kaplan-Meier curves of survival versus reamputation: presence or absence of ankle dorsi exion limitation.