Recurrent thrombosis as a clinical presentation of Whipple disease

Abstract

Whipple’s disease (WD) is a rare infectious disease with a wide clinical spectrum. Associated thrombotic manifestations are not well described in WD, only related to ‘stroke-like syndrome’. We present a case of a 39-year-old man with a 1-year history of self-limited episodes of fever, associated with generalised adenopathies and recurrent superficial and deep venous thrombosis events, which have resorted four times despite the anticoagulant treatment. Finally, the patient is diagnosed with WD. Following treatment the patient improved in his general condition, and no more episodes of fever neither thrombosis appeared during a follow-up of more than 3 years.

Background

Whipple’s disease (WD) is an infrequent infectious disease, caused by the bacterium Tropheryma whipplei. The disease was first described in 1907 by George H Whipple,¹ but the causal agent was not identified until the 1990s.² T. whipplei appears to be present in the general environment (eg, contaminated soil and sewage plant influxes),^{3 4} and this bacterium can be involved in chronic systemic infection, acute or chronic localised infections and is also well described that it can be carried asymptomatically.^{5 6}

The chronic systemic infection is the classic WD, with the typical picture of a Caucasian male between 30–50 year-old that suffer of gastrointestinal symptoms (abdominal pain and diarrhoea), joint involvement with intermittent migratory arthralgia or arthritis and important weight loss.^{3 5 7} T. whipplei can likewise cause chronic localised infections without systemic involvement, as well as blood-culture negative endocarditis, encephalitis, uveitis or lymphadenopathy.^{5 8} It has been also detected in several acute infections like pneumonia, gastroenteritis and bacteraemia, and it is postulated that the bacterium could produce the acute infection and after that evolve to a bacterial clearance, or in a few number of cases persist in the host and produce a chronic subtype.^{5 8} Because of the symptomatic variability of the disease, the diagnosis should be considered in many different clinical scenarios.³

The diagnosis has been classically based in the histological examination of duodenal biopsies, showing positive Periodic Acid–Schiff (PAS) staining.⁹ However, the diagnosis of atypical cases needs a more complex approach using other techniques, such as the quantitative real-time polymerase chain reaction (PCR) assay targeting repeated T. whipplei sequences (qPCR). Fenollar et al proposed a scheme to diagnose classic WD based on an initial screening analysing qPCR in saliva and faeces and if any results are positive perform a duodenal biopsy with PAS staining and qPCR to confirm diagnosis.⁵ In patients with suspicion of localised infection; qPCR of cerebrospinal fluid, lacrimal fluid, synovial tissue, lymph node or skin biopsies can be used, and also PAS staining of lymph nodes or other specimens.⁹ There are some authors who suggest that 18-FDG-PET may be able to contribute to the initial assessment and follow-up of this disease.⁸

When diagnosis is confirmed, treatment should be promptly initiated. For many years, the standard antibiotic treatment was considered an induction therapy with ceftriaxone (2 g/day) for 14 days followed by oral treatment with trimethoprim–sulfamethoxazole (960 mg/day). In the last years some authors have considered that the only bactericidal treatment against T. whipplei is the combination of doxycycline (200 mg/day) and hydroxychloroquine (200 mg three times/day), which is currently used as an alternative therapy. In classic WD a duration of at least 12 months of therapy is recommended, but some authors advocate more prolonged duration, especially for localised infections.^{5 7 8}

Case presentation

A 39-year-old man was admitted to our University Hospital for a lower right limb pain. At the moment of consultation, the patient was working as a sweeper and had a history of active smoking and bilateral optic atrophy due to a traffic accident.

Over the past year he required several hospitalisations for self-limited episodes of fever, which were going up to 38°C, associated with generalised adenopathies (with cervical, axilar, retroperitoneal and mesenteric location), as well as hepatomegaly and isolated episodes of abdominal, joint and chest pain with pericardial characteristics. Throughout the clinical evolution, the patient also presented episodes of both superficial and deep venous thrombosis, which have resorted up to four times despite anticoagulant treatment with acenocoumarol, maintaining 3–3.5 international normalised ratio target and second-line treatment with dabigatran.

Investigations

Within the diagnostic study multiple biochemical analyses were performed showing increased inflammatory parameters with negative results for complete serological and autoimmunity profile. An electromyography (EMG) was done for bilateral foot dysesthesia referred by the patient, revealing mild distal sensitive neuropathy. A brain MRI noticed a chronic cerebrovascular disease. As a consequence, a cardiological study was completed with transthoracic echocardiogram, that showed a preserved biventricular function with no evidence of endocarditis, as well as a Holter EKG study without evidence of arrhythmia. Owing to the repeated thrombotic phenomenon, during an inter-hospitalisation period, a study of hereditary thrombophilia was performed revealing only a protein C partial deficiency.

During the different hospitalisations, up to three PET/CT (figure 1) were performed searching for possible neoplastic/lymphoproliferative diseases revealing a variable metabolism of adenopathies, with no evidence of primary neoplasm. The study of fever of unknown origin was completed with a bronchoscopy with bronchoaspirate, and a colonoscopy that were unaltered and a fine needle aspiration biopsy of a mediastinal adenopathy that showed polyclonal, mature B lymphocytes, without immunophenotypic alterations.

Figure 1.

Total body maximum intensity projection. PET/CT showing increased uptake in lymphadenopathies of the pulmonary hilum, mediastinum and splenic hilum and multiple abdomino-pelvic normometabolic nodes. Hypermetabolic axillary, inguinal, supraclavicular and latero-cervical lymphadenopathies.

The patient was admitted to our Department for a new right iliofemoral and saphenous deep venous thrombosis (DVT) despite anticoagulation with dabigatran 110 mg/12 hours. Blood test showed microcytic and hypocritical anaemia (haemoglobin 96 g/L), with mixed characteristics of iron deficiency and chronic disorders, as well as elevation of inflammatory markers (erythrocyte sedimentation rate 120 mm/hour and C-reactive protein 20.16 mg/dL) and lactate dehydrogenase (535 IU/L) with mild hypoalbuminaemia (3.1–3.2 g/dL) and normal total proteins (7.1–7.2 g/dL). A urine dipstick test was performed which did not show proteinuria. Blood cultures were negative. A new PET/CT was performed and showed metabolic stability of adenopathies.

Diagnosis

Contemplating WD as a possible cause of the clinical picture, serum PCR for T. whipplei (TW) was determined, being positive with high load. A cervical adenopathy biopsy was also performed with histopathological outcomes, which showed non-necrotising granulomatous lymphadenitis and isolated positive PAS macrophages with positive TW PCR. Ziehl-Neelsen and methenamine silver staining were negative. To confirm the diagnosis, a duodenal biopsy was done which showed preserved villous architecture and mild chronic non-specific inflammatory infiltrate, histiocytic aggregates, absence of lymphatic ectasia and negative PAS staining. The PCR for TW was positive. PCR were also performed on faeces and saliva, being positive.

Treatment

Treatment for WD was started with endovenous ceftriaxone 2 g/day for 4 weeks and maintained by trimethoprim–sulfamethoxazole orally for more than 1 year. Low weight heparin was also initiated and continued by direct acting oral anticoagulants for the recurrent DVT treatment.

Outcome and follow-up

In the following months the patient had a great general improvement of his condition, weight gain, correction of inflammatory parameters and absence of new episodes of fever or thrombosis during a follow-up of more than 3 years (figure 2).

Figure 2.

Timeline of the different consultations and hospitalisations of our patient and evolution after antibiotic treatment. DVT, deep vein thrombosis; SVT, superficial vein thrombosis; DOAC, direct oral anticoagulant; TMP-SMX, trimethoprim–sulfamethoxazole.

Discussion

WD is an infrequent disease that can simulate haematological, rheumatological or other infectious diseases. Since WD display a vast and wide range of symptoms, and chronic evolution is usual, a high level of clinical suspicion is required to reach the diagnosis, and in our perspective, it can be commonly misdiagnosed.

In the reported case, the main symptoms were related to chronic inflammation, endothelial damage and the hypercoagulable state without other typical symptoms of the disease (recurrent abdominal pain, diarrhoea, arthralgia, etc.), and the diagnosis was delayed for 1 year despite multiple tests.

To our knowledge, only another case report revealed the association between WD and thrombosis. This patient presented bilateral iliofemoral thrombosis 2 years before the diagnosis, followed by a subacute presentation of diarrhoea, appetite loss, abdominal pain and weight loss. Endoscopies were done, revealing lymphangiectasia and villous oedema in the second duodenal portion, with microorganisms reactive to PAS in pathological examination. Treatment with antibiotic and enoxaparin was initiated and the patient improved gastrointestinal symptoms with normalisation of inflammatory activity tests and ultrasonography signs of venous thrombosis.¹⁰

We consider that the prothrombotic state of our patient was multifactorial, related to the smoking status, protein C partial deficiency and the hypercoagulability state associated to the sustained inflammation caused by the infection.

Related to the aetiology of the protein C partial deficiency, which may be caused by a protein-losing enteropathy (PLE), despite the fact that we do not have a new determination after the start of the treatment, we consider that must probably be congenital. This is based on the fact that we have no evidence of significant PLE in our patient (mild abdominal symptoms without diarrhoea, mild hypoalbuminaemia and preserved villous architecture without lymphatic ectasias in the duodenal biopsy). However, the PLE could have played an important role in the other published case.

We believe that one of the most relevant factors for thrombosis in our patient is the inflammatory state associated to the infection. Our patient had persistently elevated acute phase reactants, and it is widely described in the literature that infection and inflammation can activate the coagulation cascade by multiple mechanisms such as tissue factor-mediated thrombin generation, downregulation of physiological anticoagulant mechanisms and inhibition of fibrinolysis.¹¹ In addition, infections can cause endothelial damage due to the pro-inflammatory mediators that shift endothelial functions from the homeostatic into the defensive mode, as extensively described in the context of the COVID-19 pandemic.¹²

Taking into account these cases, we suggest that in patients with venous thromboembolic disease, without the classical risk factors (obesity, smoking, surgery, cancer, genetic thrombophilia, oral contraceptives, etc),¹³ as well as in patients with recurrent thrombotic events despite anticoagulant treatment, the diagnosis of WD should be thoroughly considered, especially if patients present fever or inflammatory parameters in laboratory tests.

Learning points.

• Whipple disease can simulate haematological, rheumatological or other infectious diseases.

• Whipple disease requires a high level of clinical suspicion to reach the diagnosis, and it can be commonly misdiagnosed.

• In patients with thrombosis without the classical risk factors, as well as in patients with recurrent thrombotic events despite anticoagulant treatment, the diagnosis of Whipple disease should be thoroughly considered.

Ethics statements

Patient consent for publication

Consent obtained directly from patient(s).