Risk Factors for Amputation in the Surgical Treatment of Hemophilic Osteoarthropathy: A 20-Year Single-Center Report

Purpose Hemophilic osteoarthropathy (HO) is a common cause of spontaneous bleeding in hemophiliacs. Surgical procedures are necessary for patients with severe HO. As a last resort, amputation is sometimes needed to treat complex HO cases. This study aimed to review the existing records of patients who underwent amputations in HO surgical treatment, summarize the risk factors, and provide relevant references for surgeons. Methods We retrospectively reviewed the records of hemophilic patients received surgeries at Peking Union Medical College Hospital between 2000 and 2020. The amputation patients without neoplasm or acute trauma were screened out. Patient information was extracted from medical records. A literature retrieval of hemophilic amputation cases was conducted via PubMed. The risk factors of amputation were summarized and analyzed via descriptive statistics and Fisher's precision probability test. Results Four male hemophilia A patients out of 201 patients underwent lower limb amputation. The reasons of amputation contained severe pseudotumor with factor inhibitor and large bone defects, recurrent pathological fracture with pseudotumor, skin ulcer with chronic osteomyelitis, and pseudotumor with systematic infection. In cases reported in retrieved literature, severe pseudotumor with complications, bacterial infection and factor inhibitor were common factors. Discussion. As the first study on amputations in HO patients, we found that severe hemophilic pseudotumor, chronic bacterial infection, and coagulation factor inhibitor was potential risk factors for amputation. Sufficient factor replacement therapy is fundamental in the prevention of amputation. The early diagnosis and specially designed surgical techniques could improve the rate of limb salvage.


Introduction
Hemophilia comprises hereditary hemorrhagic disorders characterized by coagulation system deficiencies. e most common cause of hemophilia is coagulation factor VIII or IX deficiency, causing hemophilia A or B, respectively [1]. Coagulation factor deficiencies lead to spontaneous intra-articular or intramuscular bleeding. e persisting chronic stimulation of blood stasis causes musculoskeletal lesions called hemophilic osteopathic medicine (HO). Recurrent hemarthroses consequent to bleedings contributed to cartilage degeneration named hemophilic arthritis. [2]. Inappropriate treatment of intramuscular bleeding may cause soft tissue hematoma called hemophilic pseudotumor [3].
Hemophiliacs also have a higher prevalence of osteoporosis and pathological fractures [4,5]. HO is associated with serious pain and motion dysfunction, especially in hemophiliacs administrated with limited hemostatic agents during childhood, and required surgical interventions for pain relief [6,7].
With the developments of coagulation factor replacement therapy and perioperative management, hemophiliacs can receive safe surgical procedures. Total joint arthroplasty is used to treat end-stage hemophilic arthritis [8]. Surgical resection is applied in treating hemophilic pseudotumor [9]. In the early years, amputation of extremities was used in severe HO cases due to imperfect surgical technology [10]. Unfortunately, there are still amputation cases reported in modern literature, with no systematic summary for amputation risk factors and prevention strategies [9,11,12]. In this study, we retrospectively evaluated amputation cases in the surgical treatment of HO, summarizing their clinical features and risk factors. As the first study on amputations in hemophilia patients, we hope that the result can provide a reference for surgeons facing the hard choice of limb salvage or amputation.

Materials and Methods
We retrospectively reviewed the medical records of hemophiliacs who underwent surgery for HO at Peking Union Medical College Hospital between 2000 and 2020. is study was approved by the appropriate institutional ethics review board. e patients who received amputation of extremity or joint disarticulation were screened out. Patients were excluded if the amputation was performed due to nonhemophilic diseases such as neoplasm or acute trauma. We recorded the following data: demographic data, type and severity of hemophilia, presence of factor inhibitor, aetiological antecedent, clinical preferences, and management. For all included amputation patients, follow-ups were performed by telephone, Internet, or interviews.
Because amputation is rare in hemophiliacs, we conducted literature retrieval using PubMed. e search terms contained "hemophilia," "amputation," and "disarticulation". e publication type was not limited. Clinical characteristics and amputation reasons of all reported cases were also recorded. e results about amputation cases were shown in descriptive statistics and Fisher's precision probability test. Summary statistics were performed using Microsoft Excel 2016 (Microsoft, Redmond, WA) and SPSS Statistics 19.0 (SPSS Inc, Chicago, IL, US). Numerical data were summarized in tables as arithmetic means ± standard deviation. Nominal and ordinal data are shown as frequencies with percentages. e level of significance was defined at p � 0.05.

General Characteristics.
In total, 201 patients with 262 surgical procedures were included. Four patients underwent four amputation surgeries, with a prevalence of 1.99% (4/ 201). None of the patients suffered neoplasm or acute trauma, so no case was excluded (Table 1). All four amputation patients were male with hemophilia A: two of them had received regular factor replacement before and two only received factor replacement on demand. Factor inhibitor was found in one patient. Only one case directly received amputation rather than as a secondary procedure. From the literature, we identified 27 amputation cases from 14 papers published between 1992 and 2021. e majority of reported amputation cases had hemophilia A (23/27, 85.19%). Inhibitors were identified in seven hemophilia A cases and one hemophilia B case ( Table 2).

Reasons for Amputation. Reasons for amputation varied among patients.
e reasons for four amputations were variable ( Table 2). Patient 1 accidently broke his left hip five years after bilateral total hip arthroplasty (THA). In the local hospital, he was diagnosed with a left hip periprosthetic fracture, received factor VIII infusion with limb braking for two months. e fracture did not heal, and factor inhibitor (6.8 BU/ml) was the first time found for him. After six months' termination of factor replacement, the activity of the inhibitor decreased to 1.9 BU/ml, but the left thigh swelled significantly, which was a large pseudotumor with severe bone defects in radiograph ( Figure 1). No conventional surgical technique except amputation could fix it. Patient 2 suffered a pathological fracture of the right femur caused by trauma. He received five months' bone traction at a local hospital. Two years later, a pseudotumor with bone defects was identified in the right inner thigh. In the next four years, he received twice resection of pseudotumor and one open reduction with allograft bone graft fixation of pathological fracture in the defected site. However, the fixation failed with the recurrence of pseudotumor. Methicillin-resistant Staphylococcus epidermidis was identified within wound secretion. e fracture, pseudotumor, and infection led to amputation (Figure 2(c)).
Patients 3 and 4 lost walking ability for years because of HO in knees. Patient 3 received regular factor replacement and developed two 3 cm ulcerative wounds on the right proximal tibia in the third year on the hospital bed. A surgical debridement attempt failed. Twelve years later, his right knee was fixed at a 90 angle ( Figure 3). X-ray findings suggested a sinus between the wound and tibia bone marrow, indicating suspected chronic osteomyelitis ( Figure 4). A similar 1 cm abrasion was found on the right inner thigh of Patient 4 after minor seizures for six years on the hospital bed. Two years later, the whole right lower limb began swelling and bleeding. Cultivation of wound secretions revealed the presence of Pseudomonas aeruginosa. Systematic application of antibiotics failed to control the persistent fever (37.5°C). Chronic infections were responsible for their amputations.
Among these four patients, the most common reasons for amputation were pseudotumor (3/4, 75%) and bacterial infection (3/4, 75%). e statistical analysis showed a significant difference of these factors between amputation and nonamputation cases (p � 0.001). In cases reported in literature, hemophilic pseudotumor (24/27, 88.89%) and bacterial infection (14/27, 51.85%) were also common factors. ere was only one amputation patient carried factor inhibitor in our four patients. However, he was the only inhibitor carrier among included patients. e previous amputation cases contained seven cases carrying inhibitor (25.93%), indicating that inhibitor could be another potential risk factor (Table 2).

Perioperative Treatment and Follow-Up.
All amputation patients underwent coagulation factor replacement during perioperative period under the guidance of hematologists. Two patients received aboveknee amputation while hip disarticulation and high-thigh amputation were separately performed on the other two patients. e last follow-up assessment was conducted in June 2021 (Table S1). Patient 1   was the only patient suffered from severe postoperative complications. He used recombinant human factor VIIa (rhFVIIa) and human prothrombin complex concentrate (PCC) for perioperative replacement. e FVIII: I level was beyond the limit (>64 BU/ml) with progressive anemia (96 to 43 g/l) and increased wound drainage (40 to 230 ml/d); thus, factor replacement was halted. In the next 63 days, he underwent three surgical debridement under intermittent factor replacement. Consequently, the inhibitor level decreased to 35.2 BU/ml. Staphylococcus epidermidis was once detected in wound secretions, so vancomycin was systematically applied. He was transferred to a local hospital 68 days  Pain Research and Management after amputation, and the incision healed completely after another 40 days of regular dressing. Patient 4 was transferred to the intensive care unit after surgery for three days due to a systematic infection. He recovered well and was discharged after the removal of stitches. Two years later, he died from a major epileptic event, which was found before the abrasion of the knee. In the last follow-up, all remaining patients reported no complaints, partly recovered self-care ability with crutches ( Figure 5).

Discussion
At present, amputations are mainly used in chronic nonhealing ulcers or severe infections [13,14]. is paralytic operation reduces life quality, so limb salvage is usually the first choice [15]. ough significant progress has been recorded in surgical techniques for hemophiliacs, amputations are still the final option for some severe HO cases. However, to our knowledge, no report has specially investigated the reasons for amputation in hemophiliacs. Due to the low incidence, these amputations have only been reported in isolated case reports or studies with different aims (Table 3). e present study described four amputation cases of hemophilia patients who underwent surgical treatment in a single medical center with an average follow-up of 2 years.
Severe hemophilic pseudotumor was seen in three of our four cases, which is also common in previously reported amputation cases. Recurrent spontaneous bleeding promotes the enlargement of the hematoma, resulting in incredible limb swelling and disability. e circumference of the lower limb could be several times larger than the normal value (Figure 1(c)). In extreme cases, pseudotumor compression can cause severe compartmental syndrome [21]. Pathological fracture or osteolysis is seen in half of pseudotumor amputations (2/4 in the present study and 12/24 in previous reports). e fracture is often the initial pseudotumor cause or complication. Osteoporosis is a potential explanation for these fractures. e prevalence of bone mineral density (BMD) reduction can be up to 50%-80% in hemophiliacs [4]. Systematic analyses have suggested a relationship between low BMD and severe hemophilia, but precise mechanisms linking the conditions remain unclear [28,29]. Osteolysis is mainly seen in patients with massive or multiple pseudotumor. e loss of BMD is accompanied by shape deformation. In X-ray imaging, the affected bone has a characteristic bubble-like shape and the bone will finally disappear in the pseudotumor. e exact mechanism of the osteolysis remains unclear; inflammatory reactions during hematoma absorption may induce bone resorption. As a result of fracture and osteolysis, there is a high risk of recurrent internal/external fixation failure and fracture nonunion, leading to amputation. For patients with huge osteolytic defects, total femur replacement or custom-made prosthesis is applied in some limb salvage cases, but the longterm outcomes require further observation [30,31]. Surgeons should be aware of unexplained swelling of extremities in hemophiliacs, especially the patients with a history of trauma or poor factor replacement. Adequate factor replacement is the first step in preventing bleeding and controlling pseudotumor. Based on the experience of Patient 1 in our study, we suggest regular radiology checks for osteolysis spaced at a time interval of no more than six months for patients with pseudotumor. e earlier discovery of osteolysis means a higher possibility of limb salvage. Perioperative and postoperative factor replacements are key factors affecting the success of surgical procedures in hemophiliacs; bleeding owing to inadequate replacement worsens the healing situation, and the acute hematoma may cause thrombosis via compression [27].
In the present study, the relationship between infection and amputation was highlighted. Abrasions or small wounds are common on swollen limbs with pseudotumor. Coagulation disorders make curating these wounds very hard, leaving them at high risk of chronic bacterial infection. e surgical removal of infection foci is a must for these patients. Control of coagulation disorder is the fundamental infection prevention solution. e replacement of infected joints is also a high-risk factor for amputation. A retrospective survey of the Danish population identified prosthetic joint infections as the most common cause of limb amputation after primary total knee arthroplasty [32]. Active cutaneous or deep tissue infections are important risk factors for prosthetic infections [33]. ere are also reports of amputations in hemophiliacs caused by uncontrolled infection after knee   [22]. So, patients with chronic infection or unhealed wounds are not suitable for joint arthroplasty. Arthrodesis may be a safer choice to avoid amputation. In all hemophiliacs included in the current study, Patient 1 was the only inhibitor carrier. Meanwhile, about a quarter of reported amputation cases had inhibitors, most of which underwent amputation as the first surgical treatment. It seems that hemophiliacs with inhibitors had higher possibility of amputation. e existence of inhibitors, antifactor antibodies, decreased the activity of coagulation factors, making hemostasis management difficult [34]. e etiology of developing factor inhibitors remains incompletely understood. Current evidence suggests a potential relationship between environmental factors (surgery, severe bleeds, and infections) and factor inhibitors [35]. Bypass agents such as rhFVIIa and PCC have adequate hemostatic effects and low rates of complications in surgical procedures involving inhibitor-positive patients [36][37][38]. Sufficient factor replacement plays a more important role in the perioperative management of hemophiliacs with factor inhibitors. Patient 1 in the present study was switched from rhFVIIa to PCC only three days after surgery for economic reasons. en, he suffered from bleeding and incision complications. We advise that factor replacement via bypass agents should be extended.
ere are some other risk factors for amputation. Most amputation patients are diagnosed with moderate or severe hemophilia A, indicating the influence of hemophilia type and severity. However, the prevalence of hemophilia A is about six times higher than that of hemophilia B, which may affect the incidence of severe HO and amputation. Socioeconomic factors also influence the amputation rate [1]. As we discussed above, sufficient factor replacement is important to prevent amputation. It is reported that 51% of people in high-and upper-middle-income countries use 94% and 92% of the total global quantity of factors VIII and IX, respectively [39]. For patients in developing countries, like Patient 2 in our study, regular replacement is still too expensive. Megaprothesis for osteolysis is also costly; patient 1 in our study could not afford this. In the future, gene therapy might completely eradicate this hereditary disease, but economic issues will still be barriers for many patients [40].
e retrospective nature of this study is a limitation. Furthermore, this study involved a heterogeneous group, a small number of cases, and lacks a randomized control group. Considering that amputation is rare in hemophilia treatment, a multicenter review may provide additional evidence and enhance the meaningfulness of the result.
In general, amputation of the extremities is the last but necessary option for patients with severe HO. Severe hemophilic pseudotumor, chronic bacterial infection, and factor inhibitors are threatening risk factors for amputation in hemophiliacs. Sufficient factor replacement therapy is fundamental in the prevention of amputation. Early diagnosis of severe complications such as osteolysis and special surgical techniques such as total femur replacement could be useful in limb salvage.

Data Availability
e datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethical Approval
is retrospective study was approved by the Institutional Review Board at Peking Union Medical College Hospital with protocol number S-K1737.

Conflicts of Interest
e authors declare no conflicts of interest.