Leucocytoclastic vasculitis due to acute bacterial endocarditis resolves with antibiotics

Abstract

Infective endocarditis is associated with a variety of clinical signs, but its association with multisystem vasculitis is rarely reported. A high index of suspicion is necessary to differentiate a primary autoimmune vasculitis from an infectious cause as the wrong treatment can lead to significant morbidity and mortality. We present a 71-year-old female patient with negative blood cultures, on antibiotics for recent bacteraemia, who presented with cutaneous and renal leucocytoclastic vasculitis. Workup revealed a vegetation adjacent to her right atrial pacemaker lead consistent with infective endocarditis and her vasculitis completely resolved with appropriate antibiotics.

Background

Infective endocarditis (IE) can cause multisystem vasculitis due to immune complex deposition or by inducing autoantibody formation.^1–9 Immune suppression is the standard of care for autoimmune vasculitis, but if an underlying infectious cause is present it can be detrimental. Therefore, a high index of suspicion is necessary to differentiate a primary autoimmune vasculitis from an infectious cause.

Case presentation

A 71-year-old woman with history of sick sinus syndrome, permanent pacemaker implantation, and treated methicillin sensitive Staphylococcus aureus (MSSA) bacteraemia was transferred to our facility due to new onset cutaneous lesions, acute renal failure, and acute hypercapnic respiratory failure due to pulmonary oedema and obesity-related hypoventilation.

She had been diagnosed with MSSA bacteraemia 3 weeks prior and was previously receiving intravenous nafcillin at a skilled nursing facility prior to hospitalisation. At the time of initial diagnosis, transesophageal echocardiogram (TEE) was negative for IE and repeat blood cultures were consistently negative. Despite negative blood cultures, a high index of suspicion for bacterial infection had warranted her treatment with vancomycin and cefepime in the week prior to her transfer of care.

On presentation, she was mechanically ventilated and mildly sedated. Cardiac exam revealed no murmurs. She had a painless, non-blanching, purpuric rash with scattered serous bullae concentrated on her distal upper extremities (figure 1). Initial lab results were significant for white cell count of 10.4×10⁹/L with 85.7% neutrophils, platelet count of 234×10⁹/L, blood urea nitrogen of 66 mg/dL, and creatinine of 3.46 mg/dL. Chest X-ray revealed bibasilar opacities and nodular densities in the right lung. Follow-up CT chest determined the nodules were sequelae of prior granulomatous disease, and there were no cavitary lesions. ECG showed an atrially and ventricularly paced rhythm without T wave changes. Blood, sputum and urine cultures did not grow bacteria, though procalcitonin on admission was 2.51 ng/mL.

Figure 1.

Top—Non-blanching purpura of right palmar hand and small punctate lesions. Bottom—scattered serous bullae of the proximal left upper extremity.

Investigations

Disseminated intravascular coagulation (DIC) was ruled out with a D-dimer of 5.9 µg(fibrinogen equivalent units)/mL, fibrinogen of 526 mg/dL, and the absence of schistocytes on peripheral smear. Furthermore, protein C activity was 145% and protein S of 81%.

Microscopic urinalysis for oliguric renal failure was positive for blood and spot urine studies indicated non-nephrotic proteinuria of 1.32 g/day. Anti-nuclear antibody was negative and complement levels were within normal ranges with C3 of 102 mg/dL and C4 of 52 mg/dL.

C-ANCA and P-ANCA antibodies were <1:20, anti-MPO and anti-proteinase 3 (PR3) antibodies were not detected, but atypical P-ANCA antibodies were elevated in serum at >1:20. Antiglomerular, anti-PLA2R and cryoglobulins were negative.

Biopsy of a purpuric lesion revealed a neutrophil-predominant inflammatory infiltrate of the small vessels and extravasated red blood cells consistent with vasculitis (figure 2). Special stains including periodic acid-Schiff for fungi, Grocott’s methenamine silver and Gram stain were negative for microorganisms.

Figure 2.

Left—H&E stained section of a skin punch biopsy specimen with blue-inked margin (2×). Acute inflammation is present within the superficial and deep dermis. Right—Vessels with transmural neutrophilic infiltrate and karyorrhexis consistent with vasculitis (20×). Extravasated red blood cells are also present. (Images provided by Dr. Merryl Terry, Department of Pathology and Laboratory Medicine, University of Louisville School of Medicine, 2020.)

TEE was obtained and revealed a left ventricular ejection fraction of 70% with normal intracavitary size and a normal right ventricular size without strain, as well as a 1.3×0.9 cm echogenic density at the base of the tricuspid valve, in near proximity to the traversing pacer lead, consistent with a bacterial vegetation (figure 3).

Figure 3.

Transesophageal echocardiogram revealing a 1.3×0.9 cm echogenic density in the right atrium at the base of the tricuspid valve in proximity of a right ventricular transversing pacemaker lead.

Differential diagnosis

Purpura fulminans (PF) is a life-threatening syndrome involving DIC and endovascular thrombosis.¹⁰ It can manifest during or after severe infections usually as the result of endotoxin-producing bacteria or autoantibodies against protein C or S. Rarely, it is caused by inherited protein C or S deficiencies. Hallmark signs are erythematous macules on the trunk and extremities which rapidly progress to indurated, nonblanching, and advancing lesions. Without treatment, this ultimately results in full-thickness skin necrosis and haemorrhagic bullae. Circulating anticoagulants, protein C, protein S and antithrombin are depleted, and treatment is aimed at repletion, namely with protein C concentrate and anticoagulation.¹⁰ Our patient’s presentation raised concern for PF. She appeared septic and had haemorrhagic lesions, erythematous macules and serous bullae on her proximal and distal extremities. Lab work, including normal protein C and S activity and high fibrinogen, lowered our suspicion. Skin biopsy definitively ruled out PF as there were no endovascular thrombi seen.

Henoch-Schonlein purpura (HSP) is a disease of deposition primarily affecting children that can be incited by respiratory infections or medications. Approximately 10% of patients are adults, in which case there is a correlation with underlying malignancy.¹¹ HSP presents clinically with cutaneous palpable purpura, renal involvement, arthralgias and arthritis, and gastrointestinal bleeding. Histopathologically, HSP resembles our patient’s biopsy with leucocytoclastic vasculitis (LCV) of the small superficial vessels, infiltration by neutrophil granulocytes, and extravasated erythrocytes. Immunofluorescence will reveal IgA and C3 deposition in the walls of blood vessels.¹¹ Although usually self-limited, severe cases can progress to renal failure, and if nephritis or nephrotic syndrome is present, treatment is warranted. Steroids are commonly used, but if there is renal involvement, mycophenolate mofetil, cyclosporine, or rituximab may be better therapies.¹¹ This patient’s history and lab findings were not suggestive of HSP.

The disease category of C3 glomerulopathies includes C3 glomerulonephritis and dense deposit disease. The morphological pattern of injury is termed membranoproliferative glomerulonephritis, characterised by capillary wall thickening and mesangial enlargement on light microscopy due to predominately subendothelial deposition of immunoglobulin and/or complement and mesangial proliferation. The underlying cause can be due to antibody–antigen complex deposition due to autoimmune disease or chronic infection, or due to mixed cryoglobulinemia with IgM–IgG complexes.¹² Clinically, patients present with haematuria, variable proteinuria, and can progress to renal failure with rapidly progressive glomerulonephritis. Extrarenal manifestations are rare and are typically limited to ocular drusen and acquired partial lipodystrophy.^{13 14} Our patients cutaneous lesions were not suggestive of a C3 glomerulopathy and cryoglobulins were negative in serum.

Drug hypersensitivity reactions (DHR) can be separated into two categories. Type A reactions are common, predictable reactions owing to the pharmacological properties of a specific drug and are the majority of reactions.¹⁵ Type B DHR result from immune reaction or stimulation and not the pharmacological drug properties.¹⁵ Before transfer to our facility, our patient had been treated with multiple, short courses of antibiotics. While DHR can result form short or single courses or medicines, these are due to type I hypersensitivities and are IgE mediated. Type III hypersensitivities resulting from DHR are delayed in onset and due to prolonged, high dose pharmacologic therapy only, usually presenting after 7–21 days of medication use.¹⁵ Non-steroidal anti-inflammatory medications and penicillins are implicated in some cases, and there are rare reports of LCV due to nafcillin use.⁸ While this cannot be completely ruled out, our patient’s vasculitis and cutaneous lesions developed after the cessation of nafcillin while on less than a week’s duration of vancomycin and cefepime. Furthermore, atypical P-ANCA was positive and the lesions and nephritis resolved after treatment of the IE.

LCV due to MSSA IE was suspected based on the history of infection with various antibiotic agent use, positive atypical P-ANCA antibodies, tricuspid valve vegetation, and skin biopsy revealing neutrophilic infiltrate of small blood vessels. The diagnosis was confirmed post-hoc based on the response to appropriate treatment and resolution of symptoms with resolution of IE. Antibody–antigen complex deposition may have contributed and cannot be ruled out without direct immunofluorescence of the biopsy, but complement levels were within normal limits suggesting this was less likely. Therefore, the vasculitis was likely directly attributed to atypical anti-neutrophilic antibodies.

Treatment

On admission, vancomycin and cefepime were discontinued in favour of ceftriaxone. After discovery of the tricuspid valve vegetation, rifampin was added for biofilm penetration due to suspicion for bacterial seeding of the right atrial pacer lead. Removal of the pacemaker was considered, but as she was completely pacemaker dependent, it was reserved as a last resort. Haemodialysis was required three times over the course of 7 days for volume overload and uraemia, after which she was able to be extubated.

Outcome and follow-up

One week after the addition of rifampin, kidney function improved and there was evidence of cutaneous healing. At discharge, 17 days after admission, creatinine had improved to 1.84 mg/dL from a peak of 4.88 mg/dL and there was progression of cutaneous healing. A repeat TEE was done after 6 weeks showing resolution of the vegetation. At that time, her cutaneous lesions had completely resolved. Three months later, she suffered from a MSSA infected knee prosthetic requiring multiple revisions but had no recurrence of IE, cutaneous vasculitis, or renal vasculitis. Kidney function 1 year later had returned to baseline creatinine of 0.66 mg/dL.

Discussion

LCV is a rare but documented presentation of acute bacterial endocarditis.^1–7 Because immune suppression to treat a primary autoimmune vasculitis is potentially deadly if an underlying infection exists, physicians should rule out infectious disease, specifically IE, in high-risk patients.

Various cases have been described where workup for antibody positive vasculitis revealed underlying IE. El Chami et al described a 41-year-old woman with recent surgical wound infection who presented with a cutaneous vasculitis and urine positive for blood causing concern for HSP. She was later determined to be suffering from IE of the aortic valve and tested positive for anti-PR3 antibodies.³

Other cases illustrate the dangers of misdiagnosis and inappropriate treatment with immune suppression. Wang et al described an 87-year-old man who presented with fatigue and fever. He was found to have elevated anti-MPO antibodies and was treated with steroid pulse therapy. Despite an initial clinical improvement, his condition deteriorated rapidly and ultimately resulted in death. The autopsy revealed infective vegetations on his aortic valve with septic emboli and microabscesses in other organs.⁴

Most importantly, patients presenting with vasculitis and underlying IE who are treated with valve replacement and/or antibiotics have resolution of their vasculitis, both renal and cutaneous, and negative ANCA antibody titers afterwards. Fukasawa et al described an elderly male with rapidly progressing glomerulonephritis, bacterial endocarditis due to Enterococcus faecalis, and positive anti-PR3 antibodies. After valve replacement and antibiotics, his glomerulonephritis improved and PR3 antibodies were undetectable.⁵ Our patient also had improvement of renal function and skin findings and no longer required dialysis after 1 week of adequate antibiotics.

It may not be feasible, however, to rule out infection in every case of cutaneous vasculitis. Loricera et al assessed the presentation of cutaneous vasculitis with the incidence of an underlying bacterial cause. Of the 766 patients in their study that presented with cutaneous vasculitis, 27 were diagnosed with an underlying bacterial infection and 6 were diagnosed with underlying IE. The incidence of cutaneous vasculitis with IE was 0.8%.⁶

Rather than rule out infectious cause in every case of cutaneous of multisystem vasculitis, we recommend infectious workup in patients with risk factors such as implanted devices, mechanical valves, recent oral surgery or infected surgical site, or in patients with known bacteraemia. Our patient had both known bacteraemia and an implanted cardiac device.

The association of acute endocarditis with positive ANCA could be explained by superantigenicity in the setting of an overwhelming, nonspecific B-cell activation and immune response. Veerappan et al hypothesised that autoimmunisation occurs after massive release of PR3 from neutrophils fighting an infection.⁷ This theory may hold true for other neutrophilic antigens that play a role in the early innate immune response.

Cutaneous vasculitis is a rare but documented presentation of IE. Accompanying symptoms such as renal insufficiency and haematuria may confound the workup and diagnostic approach. Immune suppression to treat a primary autoimmune vasculitis is potentially fatal if an underlying infection exists, while treatment of the infection results in resolution of vasculitis. Therefore, physicians should rule out infection, specifically IE, in high-risk patients with vasculitis prior to initiating immunosuppression.

Learning points.

• Culture-negative bacterial endocarditis and other infections can cause secondary vasculitis.

• Vasculitis due to underlying infection mimics other disease processes such as purpura fulminans, Henoch-Schonlein purpura and drug hypersensitivity reactions.

• Vasculitis due to infection and the associated antibodies will resolve with treatment of the underlying infection.

• Improper immune suppression can result in significant morbidity and mortality.

• Physicians should rule out infection and infectious endocarditis in patients with risk factors.