Whipple’s disease: misdiagnosed as sarcoidosis with further tricuspid valve endocarditis and pulmonary embolism – a case report

Abstract

GH Whipple described a 36-year-old physician in 1907 with gradual loss of weight and strength, stools consisting chiefly of neutral fat and fatty acids, indefinite abdominal signs and a peculiar multiple arthritis. The patient died of this progressive illness. Whipple called it intestinal lipodystrophy since he observed accumulation of large masses of neutral fats and fatty acids in the lymph spaces. It was renamed Whipple’s disease in 1949. An infectious aetiology was suspected as early as Whipple’s initial report. However, successful treatment with antibiotics was not reported until 1952, which resulted in dramatic clinical responses. The cause is now known to be Tropheryma whipplei. Light and electron microscopy of infected tissue identified a gram-positive, non-acid-fast, periodic acid-Schiff (PAS) positive bacillus with a characteristic trilamellar plasma membrane resembling that of gram-negative bacteria. Whipple’s disease is extremely rare. It is a systemic infectious disorder affecting mostly middle-aged white men. The clinical presentation is often non-specific, which may make its diagnosis difficult. The four cardinal clinical manifestations are arthralgias, weight loss, diarrhoea and abdominal pain. The frequently vague articular symptoms can precede the diagnosis of Whipple’s disease by an average of 6–8 years. Lymph nodes and other tissues may present diagnostic problems, since the changes in routinely stained sections may mimic those of sarcoidosis. The detection of PAS-positive histiocytes in the small intestine remains the mainstay of the diagnosis, although Whipple’s disease without gastrointestinal involvement is described. We illustrate a case in which, retrospectively, the clinical presentation would have been typical for Whipple’s disease. However, the clinical presentation and the histological examinations of lymph nodes, liver biopsies and ascites initially were misinterpreted as sarcoidosis with consecutive immunosuppressive therapy and progressive worsening of the patient’s health presenting at least as sepsis with endocarditis.

BACKGROUND

Whipple’s disease is a rare chronic, multisystemic, infective illness caused by a gram-positive bacillus named Tropheryma whipplei. In 1907 GH Whipple described a 36-year-old physician with gradual loss of weight and strength, stools consisting chiefly of neutral fat and fatty acids, indefinite abdominal signs and a peculiar multiple arthritis.¹ These clinical features are found in most cases of the disease with arthropathy preceding the diagnosis by an average of 6–8 years.^1–4 Gastrointestinal symptoms, diarrhoea, weight loss, and concomitant anorexia lead to the typical full picture of a malabsorption syndrome. Peripheral and abdominal lymphadenopathy are also common features. A mesenteric lymphadenopathy is often identified on ultrasound but could also be present as an abdominal mass. Some patients present without gastrointestinal manifestations. The clinical presentation may vary considerably from case to case because all major organs are susceptible to infection with T whipplei. Light and electron microscopy of infected tissue identifies a gram-positive, non-acid-fast, periodic acid-Schiff (PAS) positive bacillus with a characteristic trilamellar plasma membrane resembling that of a gram-negative bacteria. Application of polymerase chain reaction (PCR) analysis demonstrates a unique bacterial 16S ribosomal RNA in the intestinal and lymphatic tissue of patients with Whipple’s disease.^5,6 Despite the presumed ubiquitous presence of T whipplei, the disorder has a striking predilection for white middle-aged European males, suggesting an underlying genetic predisposition that leads to colonisation of T whipplei throughout the intestinal tract, lymphoreticular system and central nervous system on exposure to soil microbes. All of the infected sites show a remarkable lack of inflammatory response to the bacillus. In addition, the organism exerts no visible cytotoxic effects on host cells, thereby allowing massive accumulation of the bacillus at these locations. The defective immune responses have led many investigators to implicate host immune deficiency as a predisposing factor for the disease and possibly secondary immune down-regulation induced by the bacteria. The presumed immunological defect is likely to be subtle and quite specific for T whipplei since patients are not generally predisposed to infection with other organisms. Only a few case reports have pointed to the possibility that Whipple’s disease also occurs in a setting of immunodeficiency, immunosuppression, or concomitantly with other infections.² However, the possibility of a defect in host defence predisposing to Whipple’s disease has attracted much interest in the absence of evidence of a major immune defect from the patients investigated to date.⁷

CASE PRESENTATION

We illustrate a case in which, retrospectively, the clinical presentation would have been typical for Whipple’s disease. However, the clinical presentation and the histological examinations of lymph nodes, liver biopsies and ascites initially were misinterpreted as sarcoidosis with consecutive immunosuppressive therapy and progressive worsening of the patient’s health presenting at least as sepsis.

A middle-aged farmer consulted a rheumatologist because of arthralgias of the large joints for the last 6 years and 20 kg weight loss. Laboratory investigations showed a C-reactive protein of 35 mg/l (normal laboratory values, normal range 0–5 mg/l), a hypochromic, microcytic anaemia with a haematocrit of 32% (43–50%), ferritin 237 μg/l (35–217), total protein 5.4 g/dl (6.5–8.3), albumin 2.3 g/dl (3.5–5.5) and cholesterol 92 mg/dl (75–200). Antinuclear antibodies and rheumatoid factor were negative. Because of a deep vein thrombosis of the left leg and the history of weight loss, the patient was admitted to a local hospital for additional diagnostic procedures. The radiograph and the computed tomographic (CT) scan of the chest were normal. A CT scan of the abdomen revealed increased mesenteric and retroperitoneal lymph nodes, ascites and a hepatomegaly with disseminated granulomas in the liver. Erosive gastritis was diagnosed by endoscopy. As a consequence of the anaemia, the enlarged lymph nodes and the granulomatous liver disease, a bone marrow aspiration and a surgical resection of an inguinal lymph node were performed. Histology of the bone marrow and of the inguinal lymph node showed well formed non-caseating granulomas with a sarcoid-like reaction. The Ziehl-Neelson stain was negative. At laparoscopy, biopsies of liver and peritoneum, and ascites were obtained. Histological work-up identified a granulomatous hepatitis but no malignant cells within the ascitic fluid or the peritoneum. Because of an elevated angiotensin-converting enzyme (ACE) of 123 U/l (normal laboratory values, normal range 8–52 U/l), no confirmation of an infectious disease, and the non-caseating granulomas with a sarcoid-like reaction, sarcoidosis was diagnosed. Oral anticoagulation was started for a duration of 6 months and an immunosuppressive therapy with prednisone and a short course of azathioprine was initiated. Meanwhile the patient developed uveitis which was interpreted as a part of the sarcoidosis in association with the eye. The uveitis responded poorly to corticosteroid therapy. Up to this point in the course of the therapy there was no pulmonary manifestation of the sarcoidosis.

In spite of immunosuppressive treatment, arthralgias and weight loss worsened, anaemia and elevated inflammatory markers persisted, and the full picture of a malabsorption syndrome developed. Two years after his first consultation with the rheumatologist, the patient was admitted to the intensive care unit of our department with the following symptoms: recurrent fever of 38–39°C, progressive dyspnea, extensive oedema of the arms and legs, hypotension with a blood pressure of 80/60 mm Hg, an extensive stomatitis aphthosis, and exsiccosis with prerenal failure. Laboratory tests showed typical signs of malabsorption, a hypochromic microcytic anaemia with a haematocrit of 27% (normal laboratory values, normal range 43–50%), MCV 75.9 fl (80–99), MCH 24.4 pg (26–34), total protein 4.2 g/dl (6.5–8.3), albumin 1.8 g/dl (3.5–5.5) and cholesterol 53 mg/dl (75–200). Renal function was impaired with a serum creatinine of 2.4 mg/dl (0.5–1.2) and C-reactive protein was elevated to 133 mg/l (0–5).

Transthoracic echocardiography revealed a tricuspid valve endocarditis with a characteristic vegetation fixed to the well-functioning valve (moderate regurgitation), and impaired left ventricular systolic function (figs 1 and 2). Pulmonary embolism as the cause of the progressive dyspnea was confirmed by spiral CT scan with multiple intravascular embolic filling defects in both upper lobes, and an embolus as a “saddle embolus” extending into the left lower lobe. There were no signs of a deep venous thrombosis in the venous ultrasound examination. Culture-negative endocarditis was treated with piperacillin and tazobactam. Under additional parenteral nutrition the patient’s condition improved dramatically and gastrointestinal endoscopy was performed. Grade IV gastroesophageal reflux disease, gastric ulcers Forrest III and ulcers in the caecum and colon ascendens were diagnosed. On admission, the patient had multiple impressive macular erythematous skin lesions affecting the cheeks, the right upper arm, forearm, both hands, the thighs and the scrotum with inflammatory papules and pustules. Skin biopsies were performed and histology showed dermal abscess formation and PAS-positive intracellular material (fig 3). Histology of duodenal biopsies of a second endoscopy confirmed PAS-positive structures (fig 4), and the PCR amplification for T whipplei was positive. Because cerebrospinal fluid was also positive for T whipplei, antibiotics were changed to intravenous ceftriaxon for 2 weeks followed by oral trimethoprim-sulfamethoxazole for 1 year. No neurological symptoms were noted. The intrahospital course was complicated due to acute cholecystitis and herniation of an umbilical hernia. A PAS-positive bacterium was confirmed within the gall bladder and the peritoneum. Three weeks later echocardiography showed a normal tricuspid valve with only mild regurgitation, without any vegetation or destruction, and the control of the CT scan was normal.

Figure 1.

Four chamber view with large vegetation on the tricuspid valve during diastole. LV, left ventricle; RA, right atrium; RV, right ventricle.

Figure 2.

Four chamber view with large vegetation on the tricuspid valve during systole. LA, left atrium; LV, left ventricle; RA, right atrium; RV, right ventricle.

Figure 3.

Skin biopsy – subcutis with PAS-positive histiocytes (magnification: ×600).

Figure 4.

PAS staining of a duodenal-biopsy specimen with numerous macrophages containing PAS-positive granules (magnification: ×600)

OUTCOME AND FOLLOW-UP

After discharge, the further course of the disease was characterised by recurrent episodes of fever, oedema of the legs, and pleural effusions in spite of continuous therapy with antibiotics, so that admission to the hospital was necessary every 12 weeks. We conclude that relapses of Whipple’s disease were responsible for repetitive hospitalisations, since no other source of fever could be found. Relapses were treated with intravenous penicillin and streptomycin. Five years later the patient is still alive with no further relapse.

DISCUSSION

This case illustrates a rare, but untreated, potentially fatal infection with T whipplei, which, when diagnosed accurately, generally can be successfully treated with antibiotics. The disease can have a chronic relapsing course, and the organism can persist in affected tissues for a long time, even with antimicrobial therapy. Relapses have been reported in as many as 17–35% of patients and in 12 of 17 patients who remained PCR positive on small bowel biopsy obtained after initial therapy.^2,3,8 It is assumed that relapses reflect incomplete eradication of the organism with initial therapy.

Histopathological or cytological analyses by means of PAS staining are the standard methods used for diagnosis of Whipple’s disease. The characteristic feature of the disease is the presence of macrophages with intracellular inclusions that react with the PAS stain. The inclusions reflect accumulations of degraded cell wall and intact bacteria. Electron microscopy has been recommended to confirm histopathological diagnoses, especially in extraintestinal sites.⁹ Pure, lifeless or cell-free culture of the organism had not been achieved until now. However, a group in France grew the organism in a human fibroblast cell line, subcultured it successfully, demonstrated characteristic morphology by electron microscopy, and raised antibodies in mice which reacted with patient serum by immunofluorescence.¹⁰ Currently, culture is undertaken only for research purposes in highly specialised laboratories.² Diagnostic PCR assays for T whipplei are increasingly being used to establish and confirm the diagnosis of Whipple’s disease.

Whipple’s disease is an infectious, chronic multisystem disorder. In its later stages, extraintestinal disease often involves the central nervous system and the heart.^3,11,12 Whipple’s disease should be suspected in patients with weight loss, diarrhoea, arthralgias and abdominal pain or in patients with arthralgias, fever and minor gastrointestinal complaints.⁶ In these cases, upper gastrointestinal endoscopy is indicated, and several biopsy specimens from the lower duodenum should be obtained because patchy disease involvement is possible. Despite the rarity of Whipple’s disease, the fact that it can occur in the absence of typical intestinal manifestations emphasises the importance of considering the diagnosis in patients with atypical presentations. In cases of suspected extraintestinal manifestations, it is advisable to obtain specimens from the affected anatomical sites, in addition to intestinal biopsy specimens. This strategy applies to initial presentations with minimal or no apparent intestinal involvement. In our case biopsies were taken from many organ sites but no one considered Whipple’s disease. In a second retrospective, histopathological work-up of the biopsy specimens after the diagnosis of Whipple’s disease in this case study, typical but rare intracellular rod shaped organisms could be detected.

Early or atypical presentations of extraintestinal Whipple’s disease can mimic features of sarcoidosis. There are a number of case reports of Whipple’s disease with sarcoidosis-like mediastinal presentation; however, in comparison to our case most of the patients had involvement of the lungs or pulmonary lymph nodes.^6,13–16 Isolated gastrointestinal involvement with retroperitoneal or mesenteric granulomatous lymphadenopathy is very uncommon in sarcoidosis and non-caseating granulomas in the liver are associated with many different disorders.¹⁷ In our case, the histological examination of an inguinal lymph node and the biopsies of the liver described non-caseating-granulomas with a sarcoid-like reaction, which was interpreted as sarcoidosis. Moreover, an elevated level of ACE, as in our case, does not rule out Whipple’s disease.¹⁸

Whipple’s disease with ocular involvement is rare but well described. In his review of 696 patients with Whipple’s disease, Dobbins found that 19 had evidence of ocular involvement.⁶ The most common ocular manifestations have included uveitis, vitritis, retinitis and retrobulbar neuritis. However, there is no previously reported case of ocular disease in the absence of marked central nervous system and gastrointestinal manifestations.

Cardiac involvement is common in Whipple’s disease and may present as pericarditis or blood culture-negative endocarditis, which is seen at autopsy in up to one half of patients with Whipple’s disease.^19,20 Usually, the mitral and aortic valves are diseased, and only in rare circumstances is the tricuspid valve affected.²¹ In many of these patients endocarditis is isolated, duodenal biopsy may be negative, and no other evidence of clinical Whipple’s disease is observed.²² In contrast to endocarditis of other causes, fever and previous valve disease are less frequently present in patients with Whipple endocarditis. The pathological features of surgically removed valves are thickened cusps (distorted and fibrotic), and vegetations of variable sizes are present on the valvular surface. Valvular tissue is very inflamed and vascularised and lacks calcification. The pattern of inflammation with prominent fibrosis is the key for formulating a diagnosis and consists of a predominantly mononuclear cell infiltrate with numerous foamy macrophages.²³ A review of 35 cases of Whipple-associated disease lends support to a presurgical antibiotic treatment regimen at least for 2–6 weeks, followed by long-term co-trimoxazole treatment, and suggests that surgical valve replacement might be a prerequisite for successful therapy.^22,24,25 However, there is no evidence that valvular surgery is always necessary to cure Whipple’s endocarditis. One report suggests that patients with Whipple’s disease who require prosthetic cardiac valves have a high risk of developing specific endocarditis.²⁶ To our knowledge, tricuspid valve endocarditis with ensuing pulmonary embolism has never been described as the leading manifestation of an infection with T whipplei. Differential diagnosis includes thromboembolic venous disease. Due to the fact that the vegetation was attached to the tricuspid valve and seen in further echocardiographic examinations, a thrombotic formation is very unlikely. Myocarditis is a rare manifestation and can present as new onset heart failure or sudden death.²⁷

Involvement of the skin, especially as hyperpigmentations particularly effecting light exposed areas not related to adrenal dysfunction, have been reported in up to one third of patients in a large series.² In one case skin lesions resembling dermatitis herpetiformis have been described.²⁸ Although an infection with T whipplei was suspected because of the impressive macular erythematous skin lesions with dermal abscess formation and PAS-positive intracellular material, in typical Whipple’s disease the most severe changes are seen in the proximal small intestine.¹ In our patient, the skin lesions led to a secondary work-up of the histopathological examinations resulting in the diagnosis of Whipple’s disease. The diagnosis has to be confirmed by PCR.

The presentation of Whipple’s disease is often non-specific, which may make its diagnosis difficult. The four cardinal clinical manifestations are arthralgias, weight loss, diarrhoea and abdominal pain. The frequently vague articular symptoms can precede the diagnosis of Whipple’s disease by an average of 6–8 years. Lymph nodes and other tissues may present diagnostic problems, since the changes in routinely stained sections may mimic those of sarcoidosis. The detection of PAS-positive histiocytes in the small intestine remains the mainstay of the diagnosis, although Whipple’s disease without gastrointestinal involvement is described. Clinical situations where Whipple’s disease should be considered, are unexplained malabsorption with systemic disease, unexplained systemic granulomatous disease resembling sarcoid, neurological disease characterised by myoclonus, dementia, and supranuclear ophthalmoplegia and unexplained uveitis.

LEARNING POINTS

• In general, arthritis precedes the diagnosis of Whipple’s disease.

• Arthropathy in combination with weight loss is suggestive of Whipple’s disease.

• Upper gastrointestinal endoscopy is indicated, and several deep duodenal biopsy specimens are recommended.

• Extraintestinal Whipple’s disease can mimic features of sarcoidosis.