Digital ischaemia secondary to adalimumab-induced antiphospholipid syndrome

Abstract

A 50-year-old woman with a history of Crohn’s disease treated with adalimumab presented with left hand pain and duskiness. Angiogram showed non-filling of the radial and digital arteries of the hand. Antiphospholipid antibody testing was positive, leading to a diagnosis of antitumour necrosis factor-induced antiphospholipid syndrome. Adalimumab was discontinued, and she was treated with the vitamin K antagonist warfarin and low-dose aspirin. Upon resolution of the antiphospholipid antibodies, she was transitioned to aspirin alone without recurrence of thrombosis.

Background

Agents antagonising tumour necrosis factor alpha (TNF-α) are widely used in the treatment of autoimmune conditions such as rheumatoid arthritis, psoriasis and inflammatory bowel disease. As their use has expanded, the appearance of anti-TNF-induced lupus and vasculitis has become a well-recognised clinical entity.^1–3 In contrast, antiphospholipid syndrome (APS) induced by anti-TNF agents has been less commonly described. Prior to this case, only a few examples of anti-TNF-induced APS have been reported in the literature in which immunological and clinical features of APS were documented.^{4 5} Both venous and arterial events have been described while under treatment with adalimumab, infliximab and etanercept.^6–8 There are also reports of patients treated with TNF-α blockade who developed antiphospholipid antibodies, although without an obviously increased risk of thrombotic events.^{9 10}

Like most systemic autoimmune conditions, the precise aetiology of APS remains obscure. To classify a patient as having APS, they should have autoantibodies (anticardiolipin, antibeta-2 glycoprotein I/β₂GPI or lupus anticoagulant) and a disease-defining event such as thrombosis or pregnancy loss.¹¹ Presumably, both genetic and environmental factors are at play in the emergence of APS. Here, we discuss the clinical course of a patient in whom the environmental trigger was very apparent and also reversible. For most patients diagnosed with APS, lifelong anticoagulation is recommended.¹² Should that be the case here?

Case presentation

A 50-year-old woman with a history of Crohn’s disease (age 35) and infliximab-induced lupus (age 47) presented to the emergency room with 2 weeks of progressive left hand pain (especially the second, third and fourth fingers) with duskiness of the second finger. Her radial and ulnar pulses were normal. Diagnosed at age 35 with Crohn’s disease and initially treated with ileocolectomy, she remained symptom free from her inflammatory bowel disease (IBD) until age 44 when she was admitted twice for increased bowel obstruction symptoms. She was initiated on infliximab with improvement in her IBD symptoms; however, soon thereafter she developed a symmetric small-joint arthritis, newly positive antinuclear antibodies (ANA; >1:2560) and modestly elevated anti-double-stranded DNA antibodies (17.1 IU/mL, normal <7.0 IU/mL). Prior to initiation of infliximab, ANA screen had been negative. Infliximab was subsequently halted, and she was treated with a steroid burst and taper. Twenty-eight months prior to her presentation, she was transitioned to adalimumab 40 mg every 2 weeks, after which arthritis symptoms resolved. The dose was escalated to 40 mg weekly 14 months prior to her presentation to achieve better control of IBD symptoms. Following this dose adjustment, she remained well controlled with regards to her Crohn’s disease and her arthritis until the time of her presentation.

Additional history revealed a one-pack-per-day smoking history. She did not have any history of previous thrombosis, miscarriage or pregnancy complications. Her active medications other than adalimumab were: bupropion, escitalopram and dexlansoprazole. On presentation, she described a 2-week history of progressive left hand pain and discolouration. Her examination was significant for palpable radial pulses bilaterally with delay in left-sided pulses. Her left second finger was dusky at the nail bed, and her second, third and fourth fingers were cool and tender to palpation. She did not have any new rashes or synovial thickening. Her cardiopulmonary examination was unremarkable. Her neurological examination revealed no sensory deficits.

Investigations

Initial laboratory testing revealed normal complete blood count, basic metabolic panel, prothrombin time, partial thromboplastin time, sedimentation rate, C-reactive protein and urinalysis. Antiphospholipid antibody testing was positive with anticardiolipin IgM of 73 MPL (normal 0–20 MPL), anti-β₂GPI IgM of 63 SMU (normal 0–20 SMU) and dilute Russell viper venom time (dRVVT) 52.7 s (ratio 1.55). ANA remained positive (>1:2560), as did antidouble-stranded DNA antibodies (32.6 IU/mL). Cryoglobulins and antineutrophil cytoplasmic antibodies were negative.

An angiogram of the left upper extremity showed non-filling of the radial artery distal to the flexor retinaculum, the digital artery of the thumb and the medial proper digital artery of the second digit. There was paucity of filling in the deep palmar arch (figure 1). The aortic arch and subclavian artery were angiographically normal. There was no evidence of corkscrewing within the small vessels. Intra-arterial nitroglycerin was administered without angiographic response.

Figure 1.

Angiogram of the left upper extremity showing non-filling of the radial artery (asterisk) distal to the flexor retinaculum. There is a paucity of filling in the deep and superficial palmar arches (blue arrowheads), especially at the level of the second and third digits. There is minimal filling of the digital arteries of the second digit (orange arrowheads). Intra-arterial nitroglycerin was administered without change in the filling pattern, suggestive of small-vessel thrombosis.

A transthoracic echocardiogram performed following the angiogram showed no evidence of valvular abnormality, mural thrombus or right-to-left shunt.

Differential diagnosis

Prior to the angiogram and laboratory testing, the differential diagnosis included early manifestations of thromboangiitis obliterans or a smoking-induced vasospastic process. Anti-TNF-induced vasculitis, lupus and APS were considered. A thromboembolic phenomenon unrelated to the anti-TNF agent was also discussed. The lack of corkscrewing vessels on the angiogram pointed away from thromboangiitis obliterans, and the rest of the angiogram and echocardiogram provided no evidence of thromboembolic phenomenon. The differential for de novo small-vessel thrombosis in the presence of positive antiphospholipid antibodies was thought to be either anti-TNF-induced APS or possibly an exacerbation of her historical anti-TNF-induced lupus leading to intra-arterial thrombosis. In the absence of symmetric small joint arthritis, active purpuric rash or other systemic signs or symptoms suggestive of active lupus or vasculitis, APS was felt to be the dominant diagnosis.

Treatment

Initial treatment included discontinuation of adalimumab 2 days after presentation to care. She was initiated on heparin and bridged to warfarin (goal international normalised ratio 2–3); she was also started on aspirin 81 mg/day and hydroxychloroquine. Her gastroenterologist initiated ustekinumab for maintenance treatment of the Crohn’s disease, and her gastrointestinal symptoms were well controlled once that medication was increased to 90 mg every 8 weeks. Sulfasalazine was later added for treatment of inflammatory bowel disease-associated arthritis.

Outcome and follow-up

In close follow-up, there was no evidence of further thrombotic events on warfarin and low-dose aspirin. Anticardiolipin and anti-β₂GPI antibodies were trended and remained positive for 18 months, though they began trending down almost immediately following discontinuation of adalimumab (figure 2). Following normalisation of these antibodies (and also normalisation of the dRVVT screening time), warfarin was discontinued and she was maintained on low-dose aspirin and hydroxychloroquine. She has now been followed for 39 months after anticoagulation was discontinued without recurrence of thrombosis or re-emergence of antiphospholipid antibodies.

Figure 2.

Antiphospholipid antibody levels trended from the time of diagnosis to most recent follow-up. Arrows indicate time at which adalimumab was halted and time at which warfarin was halted. Values reported in SMU for β₂GPI IgM (normal<20 SMU) and MPL for cardiolipin IgM (normal<20 MPL).

Discussion

Interestingly, there is some evidence that patients with APS have increased levels of TNF-α compared with controls.¹³ Additionally, experimental models of APS have shown a pathological role of TNF-α in the endothelial damage associated with obstetric APS.¹⁴ These and other similar studies have suggested a role for TNF-α blockade in the treatment of APS; however, outside of obstetric APS,¹⁵ this has not been evaluated.

The role of TNF-α in the pathogenesis of APS-associated thromboembolic disease is unknown and may differ from its role in obstetric complications. Cases like this one suggest that anti-TNF therapy should be pursued cautiously in APS given its potential to unmask antiphospholipid antibodies in some patients. While the mechanism of drug-induced APS has not been elucidated, it has been hypothesised that alterations in the permeability of the cellular phospholipid bilayer may play a role in the development of antiphospholipid antibodies.¹⁶ While the potential role of adalimumab in altering the cellular membrane is unknown, TNF-α itself may play a role in determining the milieu of the phospholipid bilayer.¹⁷ Patients who are predisposed, by other environmental or genetic factors, may develop APS in the setting of anti-TNF therapy; here, the patient’s history clearly indicated a predisposition towards autoimmune disease, given past emergence of inflammatory bowel disease, enteropathy-associated arthritis and drug-induced lupus. Further mechanistic work on this topic is clearly needed.

This case demonstrates the resolution of clinical and immunological APS after discontinuation of the offending medication, adalimumab. Despite initial uncertainty regarding the diagnosis, over time APS became increasingly likely. Other considerations such as thromboangiitis obliterans, exacerbation of anti-TNF lupus or anti-TNF-induced vasculitis were never substantiated. Specifically, the patient never developed signs or symptoms of vasculitis or symptoms of lupus as she had developed years before while taking infliximab. Despite continuing to smoke, she did not have progression of her symptoms, as would be expected in thromboangiitis obliterans.

Furthermore, this patient has now been thrombosis free for approaching 3 years since discontinuation of anticoagulation. Whether anticoagulation should be stopped in patients with APS is a question that has not been well studied. Small studies suggest that this may be safe in patients with history of venous thromboembolism or obstetric complications,^{18 19} but there is essentially no data available in patients with history of arterial thromboses (as here). Additionally, given this patient’s initially high-risk antibody profile,²⁰ it is unusual that the antibodies resolved and anticoagulation was successfully discontinued. While it is not possible to prove causation with a single report, this case highlights the value of the format to raise awareness of a potential rare but morbid or life-threatening association. This case demonstrates the development of transient APS in the setting of TNF-α blockade and suggests that these patients may not require lifelong anticoagulation if the offending agent is discontinued and antiphospholipid antibodies resolve.

Patient’s perspective.

Humira was given for my Crohn’s disease to replace Remicade. I experienced tremendous improvement with this medication. After a long period of time, I started experiencing a tingling sensation in my index finger. After about a week, I began feeling blood rushing to the tops of my fingers when I held my hand downward. I noticed a blue coloration developing in my palm. Testing confirmed blood clots and Humira was suspected as the cause. I was devastated because the Humira was working so well for my Crohn’s. My doctors were amazing and worked as a team to treat my APS and Crohn’s simultaneously. I am now symptom-free and have begun a new medication for my Crohn’s disease. Thank you so much to all my doctors!

Learning points.

• Antitumour necrosis factor (TNF) agents may induce antiphospholipid antibodies associated with arterial thrombosis.

• In rheumatoid arthritis and inflammatory bowel disease, patients on TNF-α blockers with new thromboses should potentially be evaluated for antiphospholipid syndrome (APS).

• In some cases, it may be safe to halt anticoagulation for patients with anti-TNF-induced APS if their antiphospholipid antibodies resolve.