Introduction
Warfarin is used extensively in clinical practice, from young people with deep vein thrombosis to elderly patients with atrial fibrillation. It is generally safe, although therapeutic responses vary considerably. Bleeding is the major complication, in the form of ecchymosis and purpura. However, warfarin can also cause haemorrhagic skin necrosis, maculopapular vesicular urticarial eruptions and purple digit syndrome.Reference Egred and Rodrigues1
Warfarin-induced skin necrosis was first described in 1943 by Flood, and is said to have an incidence of one in 1000 to 10 000.Reference Roche-Nagle, Robb, Ireland and Bouchier-Hayes2 It predominantly affects obese, peri-menopausal women,Reference Warkentin, Sikov and Lillicrap3 usually affecting fatty anatomical areas such as the breasts, buttocks and thighs,Reference Scandling and Walker4 as illustrated by numerous reports.Reference Amato, Berti and Fabbri5–Reference Ward and Chavalitanonda9 None of these, however, have described the condition affecting the skin of the ear.
We report an unusual case of warfarin-induced skin necrosis affecting a patient's left pinna, which posed a local management challenge.
Case report
An 82-year-old man with a history of B cell non-Hodgkin's lymphoma (in remission) and carcinoma of the prostate was referred by his general practitioner to the ENT on-call team with suspected pinna haematoma. This had originally appeared as a small, red lesion over the patient's mid-antihelix, two days prior to referral. There was no recent history of trauma or infection in the area. However, seven days prior to the current episode, the patient had been discharged from hospital after undergoing surgery for a transitional cell carcinoma of the bladder. The patient was receiving long-term warfarin therapy following three previous episodes of deep vein thrombosis. His warfarin had been stopped five days prior to surgery and recommenced with a loading dose upon discharge. The patient's international normalised ratio (INR) on discharge was within the therapeutic range.
On examination, the patient had an erythematous, swollen, indurated area of skin from the mid-concha to the lobe of his left pinna (Figure 1). There was no discharge from the ear canal or bleeding from the affected area. The affected area was cold and tender on palpation and there was no fluctuance.
Initial presentation of apparent haematoma of the left pinna.
The case was initially managed as a possible external ear infection, with intravenous and topical antibiotics. Despite this therapy, the initial discolouration failed to resolve and continued to spread to the rest of the pinna and peri-auricular area. Moreover, the oedema and erythema progressed to involve the adjoining cheek and temporal areas (Figure 2).
Progression of pinna necrosis and extension to the face and upper neck.
Other possible differential diagnoses were taken into consideration in view of the patient's past medical history, such as a paraneoplastic thromboembolic event or vasculitis. After contacting the medical team, we carried out multiple tests for vasculitides (e.g. anti-neutrophil cytoplasmic antibodies (ANCA) and cryoglobulin), autoimmune diseases (e.g. antinuclear antibodies, cardiolipin and lupus anticoagulant tests) and coagulopathies (e.g. factor V Leiden, lupus anticoagulant tests, and protein C and S). The only abnormality found was a low protein S level (66 per cent; normal level is more than 73 per cent).
A wedge biopsy of the pinna was sent for microbiological analysis, and was reported as showing no evidence of bacterial or fungal infiltration. Histopathological analysis showed areas of infarction, intravascular thrombosis, and widespread necrosis of vessel walls with nuclear debris, but no evidence of primary vasculitis (Figure 3).
Photomicrograph of wedge biopsy of left pinna, showing skin necrosis and vascular thrombosis and necrosis. Note that skin appendages, cartilage (bottom right) and one small artery (centre) are still viable (H&E; ×100).
Having exhausted other diagnoses, via a multidisciplinary approach involving plastic surgery, dermatology, microbiology, biochemistry and histopathology departments, it was concluded that this patient had warfarin-induced skin necrosis, due to rapid recommencement of therapy after his bladder surgery.
At this stage, it now appeared that the patient might require extensive debridement with resection of the pinna. Fortunately for the patient, a conservative approach was adopted and he was seen regularly in the dermatology clinic for superficial debridement and antiseptic dressings. Anticoagulation therapy was reinstated using low molecular weight heparin. The final result was that the patient lost a very small portion of his pinna without any need for more extensive surgery (Figure 4).
Left pinna following conservative debridement and dressing.
Discussion
Warfarin-induced skin necrosis is rare (occurring in 0.01–0.1 per cent of all patients on warfarin).Reference Roche-Nagle, Robb, Ireland and Bouchier-Hayes2 It usually presents in middle-aged women when the loading dose is too high. In the case of our patient, it occurred following surgery, despite the patient having received warfarin for many years, and having previously had his warfarin stopped and restarted in a similar way.
It is common for the skin surrounding these lesions to present with an accumulation of oedema in the dermis and subcutaneous tissues (peau d'orange effect). Full thickness skin necrosis presents in the late stages. Histopathological changes comprise microvascular injury with fibrin deposits in the postcapillary venules and small veins, along with haemorrhage and diffuse necrosis in the dermis and subcutaneous fat. Most specimens show a distinctive lack of vascular or perivascular inflammation, as well as a lack of arteriolar thrombosis. The differential diagnosis includes lupus anticoagulant associated skin necrosis and pyoderma gangrenosum, together with other causes of acute thrombosis associated with skin necrosis such as necrotising fasciitis and disseminated intravascular coagulopathy.Reference Sanchez-Politta, Angelillo-Scherrer, Masouye and Borradori10, Reference Chan, Valenti, Mansfield and Stansby11 In our patient, intravascular necrosis and full thickness skin necrosis were seen, without vasculitis. Cases of warfarin-induced necrosis in different anatomical sites (e.g. breast, thighs, abdomen or widespread body necrosis) have been reported to show similar appearances.Reference Warkentin, Sikov and Lillicrap3, Reference Tai, Ierardi and Alexander12–Reference Scarff, Baker, Hill and Foley15
Warfarin-induced skin necrosis is likely to be related to a temporary imbalance in the ratio of procoagulant and anticoagulant proteins, together with a rapid decrease in levels of anticoagulant protein C (occurring during the initial phase of anticoagulation), before levels of these proteins have stabilised.Reference Comp, Elrod and Karzenski16 Resistance to activated protein C has been found in about 5 per cent of the general population and in 20–60 per cent of patients with deep venous thrombosis.Reference Bauer17 Skin necrosis has been found to be associated with other hypercoagulable conditions (e.g. protein S deficiency).Reference Roche-Nagle, Robb, Ireland and Bouchier-Hayes2 Screening for protein C and S and for antithrombin III deficiency is neither sensitive nor specific in predicting the risk of developing warfarin-induced skin necrosis.Reference Chan, Valenti, Mansfield and Stansby11 Possible causes of late-onset warfarin-induced skin necrosis include: discontinuation and then recommencement of warfarin; a sudden decrease in synthetic function of the liver; and drug interactions such as sodium salicylate and ibuprofen.Reference Essex, Wynn and Jin18
The presented patient posed both diagnostic and management challenges. Diagnosis of warfarin-induced necrosis is usually delayed due to its rarity.Reference Sanchez-Politta, Angelillo-Scherrer, Masouye and Borradori10 In the current case, the site of affliction was highly unusual. Whilst warfarin-induced necrosis has been reported to affect the skin of the face, there has hitherto been no report of it affecting the pinna; it has however been noted that the trunk, face and extremities are more affected in male patients. In the present case, blood tests pointed to the diagnosis by showing decreased protein S levels, and histopathological analysis concurred by ruling out vasculitis.
In terms of management, the central issue concerned the extent to which debridement should be undertaken. At one stage, the whole pinna appeared necrotic and non-viable. However, once infection had been excluded by swabs and biopsies, a conservative approach was employed, comprising limited debridement and regular dressing. Ultimately, this approach was vindicated, resulting in preservation of most of the pinna, and saving the patient from a disfiguring deformity.
• Warfarin-induced skin necrosis is a rare adverse effect of warfarin therapy
• The condition usually affects women, and involves fatty areas such as the breasts
• This case highlights the need for a high index of suspicion in order to diagnose this condition
• In this patient, conservative management of necrosis was more appropriate than extensive excision
Otorhinolaryngological surgeons and clinicians should be aware of skin necrosis as a complication of anticoagulation therapy. Prevention of such a complication should be taken into consideration by using small loading doses on reintroduction of warfarin and close INR monitoring. Reports agree that early recognition and cessation of warfarin results in the best outcome; this prevents the development of new lesions, the need for surgical treatment or (in severe cases) sepsis resulting in death.Reference Warkentin, Sikov and Lillicrap3, Reference Scarff, Baker, Hill and Foley15 There is often a delay in recognition and therefore in treatment, so therapies such as vitamin K and monoclonal antibody purified protein C concentrates have been advocated.Reference Lewandowski and Zawilska19 Even when the condition is diagnosed early, surgical intervention may be required since the progression of the escar is independent of warfarin.Reference Chan, Valenti, Mansfield and Stansby11 A review by Chan et al. states that half the patients with this condition will need surgical intervention.Reference Chan, Valenti, Mansfield and Stansby11 However, a higher proportion of cases reported in the literature undergo surgical debridement, including amputation of the affected extremity.Reference Warkentin, Sikov and Lillicrap3, Reference Khalid13, Reference Parsi, Younger and Gallo14, Reference DeFranzo, Marasco and Argenta20 The reported case supports a more conservative surgical approach in patients whose diagnosis is not clear and who have limited, non-infected, necrotic lesions.
Conclusion
We report an unusual case of warfarin-induced skin necrosis affecting a patient's pinna, which posed diagnostic and management dilemmas. Such a presentation has not previously been reported in the literature, and highlights a pitfall for the unwary if the condition is not recognised and managed appropriately.
