Article Text

Download PDFPDF

Diagnosis and management of Whipple’s disease of the brain
  1. P K Panegyres
  1. Neurologist and Director, Neurodegenerative Disorders Research, 185 York Street, Subiaco, WA 6008, Australia; publications{at}

Statistics from

Whipple’s disease is a rare condition caused by a soil-born gram positive bacillus Tropheryma whipplei (T whipplei).1 Whipple’s disease of the brain is even rarer. It may occur in the context of a systemic illness with gastrointestinal or rheumatological features, or as a primary central nervous system infection.2 On the basis of my experience with two recent patients contrasting primary and secondary Whipple’s disease of the brain, I will suggest an approach to its diagnosis and management.


A 46-year-old woman presented with generalised tonic-clonic seizures. She subsequently developed diplopia and ataxia. Four months later she had a tremor, a fluctuating left VI nerve palsy, eye movement abnormality with impaired saccadic pursuit, and hyponatraemia secondary to inappropriate ADH secretion. Later blepharospasm, complete paralysis of upgaze, worsening ataxia and a positive Romberg’s sign appeared. She had a severe amnesic syndrome including impaired verbal and visual spatial memory. Five months after onset she rapidly deteriorated with fluctuating consciousness—her Glasgow Coma Score ranged from 3 to 14. She also experienced hallucinations and confusion.

At presentation her brain MRI had revealed atrophy and gliosis of the right hippocampus. However, concomitant with her deterioration she developed multiple nodular enhancing lesions on MRI involving the right temporal lobe, caudate nucleus, anterior commissure, right globus pallidus, left insula cortex, left hippocampus, lenticular striate territory, mesencephalon and perivascular spaces (fig 1). At this time a lumbar puncture showed an opening pressure of 17 cm of water with 464 white cells/μl (90% polymorphs, 10% lymphocytes). There were 30 red cells/μl. Cytology showed polymorphs and lymphocytes without malignant cells. The protein was 0.35 g/l (0.15–0.45); glucose was 1.6 mmol/l (concurrent blood glucose 3.1). Gram stain, India ink preparation and cultures for Mycobacteria were all negative.

Figure 1 Patient 1. (A) Axial T1-MRI, (B) axial FLAIR and (C) coronal FLAIR images reveal atrophy of the right mesial temporal lobe with persisting signal abnormality suggesting gliosis. Oedema present in the left mesial temporal lobe, caudate heads, and right putamen. (D) Coronal contrast enhanced T1-weighted image shows patchy nodular enhancement involving the mesial temporal lobes, insula cortex and lenticulostriate vessels.

She was treated with methylprednisolone, cotrimoxazole (trimethoprim-sulfamethoxazole), meropenem, frusemide, salt tablets and fluid restriction after Whipple’s disease was suspected on the basis of the development of complex eye movement abnormality and the MRI showing multiple nodular enhancing lesions (granulomatous or inflammatorydisorders including sarcoid were also possibilities). She responded to this treatment and after five days her conscious level returned to normal. PCR for Whipple’s disease on the cerebrospinal fluid (CSF) was positive on two occasions and confirmed by an independent laboratory. The PCR was also positive on the blood. There was no lymphadenopathy on systemic examination, CT of the abdomen showed no lymphadenopathy, and duodenal biopsy was negative, making systemic Whipple’s unlikely.

She was eventually able to sit out of bed, walk and manage herself independently, and was discharged home with only a mild upgaze abnormality and problems with episodic memory. She experienced hypersomnolence and snoring and was shown to have obstructive sleep apnoea on overnight polysomnogrophy, requiring an assisted ventilation device after discharge. She also had reduced sleep efficiency and fragmentation. She was able to live independently, drive, and eventually return to work as a child carer on a supervised basis. She has remained on treatment since diagnosis.

Repeat MR scan six months after discharge showed complete resolution of the nodular enhancing lesions, but with residual gliosis and atrophic change in the right hippocampus. Twelve months after her deterioration the MR scan was unchanged, with normal CSF apart from persistent oligoclonal bands which had been present on the first CSF examination and performed with concurrent blood assessment for paraproteins which was negative. (This patient has been previously published in more detail by Panegyres et al.2)


A 41-year-old man developed the abrupt onset of continuous and involuntary twitching of the muscles of the left side of his face which prevented him sleeping. He also had difficulty with his vision which he described as “blurry”, with a problem “focussing” and an inability to “track an object”, and trouble looking down. He was having problems driving. He had impaired balance. He was hypersomnolent and found it difficult to work as a metallurgist. He had been lacking in energy for some months but there had been no abdominal pain, no change in bowel habit, and no evidence of malabsorption or arthritis. There was no past or family history of neurological disease. He was a non-smoker and took no alcohol. He had been treated briefly with fluoxetine, clonazepam and ropinirole, all without benefit. His symptoms worsened.

He looked chronically ill at neurological presentation three months after the onset of symptoms. Throughout the examination he had involuntary twitching in the muscles of the left side of the face, involving left masseter and zygomaticus. Twitching also extended to the left orbicularis oris, platysma, stylohyoid and digastric muscles. There was no twitching of the muscles of the right cheek. He had occasional twitching of both frontalis muscles. He had 20 degrees of vertical upgaze and 15 degrees of vertical downgaze from the horizontal. Lateral gaze was normal. Pursuit function was intermittent, suggestive of brainstem and ocular motor nuclei involvement. There were no vertical and very reduced horizontal saccadic movements within unsuppressible blinks. Vertical gaze paresis was overcome completely by head position, indicating a supranuclear component. Accommodation reflex revealed reduced ocular fixation, and reduced convergence of both eyes with absent pupillary responses. The remainder of the neurological examination was normal, abdominal examination was also normal, and there was no arthritis.

He had two MR scans of the brain (1.5 Tesla and 3 Tesla) with and without gadolinium, both of which were normal. His blood count showed a haemoglobin of 130 g/l (135–170), a microcytosis of 81 fl (82–98), ESR 68 mm/h (1–20), platelets 428/l (140–370). White cell count was 9.6×10/l, with a normal differential (4.0–11.0). Serum albumin was 35 g/l (38–50). C-reactive protein 56 mg/l (<10). Urea, creatinine, electrolytes and liver function were all normal. CSF examination showed an opening pressure of 21.5 cm of water. No leucocytes and 10 red cells per μl. No bacteria seen. Culture negative. Protein 0.41 g/l, glucose 3.2 mml/l and concurrent blood glucose 5.2 mml/l. Cytology showed a mild CSF lymphocytosis without malignant cells. Oligoclonal bands were not detected.

There was a polyclonal increase in gamma globulins in the blood, and immunofixation electrophoresis excluded a paraprotein. Blood flow cytometry showed an abnormal population of B-lymphocytes suggestive of a B cell lymphoproliferative disorder. CT of the abdomen showed multiple para-aortic, paracaval and mesenteric lymphadenopathy, which raised the possibility of secondary Whipple’s disease of the brain. A fine needle aspiration of the abdominal lymph nodes was non-diagnostic. However, lymph node biopsy at laparotomy showed multiple collections of epithelioid histiocytes, with focal multinucleated cells, and large numbers of foamy and granular histiocytes in sheet-like arrangements. The cytoplasm of many of the cells had a granular appearance. Large numbers of Periodic Acid Schiff (PAS) positive bacilliform organisms within the histiocytes were found, diagnostic of Whipple’s disease (fig 2). Gram and Ziehl-Neelsen stains were negative.

Figure 2 Lymph node biopsy from patient 2 showing multiple collections of epithelioid histiocytes with focal multinucleated cells (arrow) and large numbers of foamy granular histiocytes. (A) and (B) H&E ×20, ×40 respectively. Considerable numbers of PAS positive bacilliform organisms identified (arrow) (C) H&E ×40. Whipple’s disease organism (arrow) confirmed by electron microscopy (D).

The presence of constitutional symptoms, eye movement abnormality and myorhythmia had all suggested Whipple’s disease, which was confirmed by PCR on CSF for T whipplei on two occasions. T whipplei was also confirmed by DNA sequencing of blood lymphocytes.

He was started on meropenem intravenously 1 g three times a day and oral cotrimoxazole two tablets twice a day. After a week the meropenem was changed to ceftriaxone 2 g intravenously per day for 14 days, for cost issues. After the first week he developed sensitivity to the cotrimoxazole and was changed to oral doxycycline (100 mg twice a day) which has been maintained continuously since June 2007. His facial twitching and eye movement abnormalities did not respond to valproate, carbamazepine, clonazepam or levetiracetam. Botulinum toxin injections were required which were effective. The eye movements, constitutional symptoms, inflammatory markers and abdominal lymphadenopathy have all improved. He has gained weight and returned to full-time work but has persistent lethargy.


Patient 1 had primary Whipple’s disease of the brain with seizures and complex eye movement abnormality without evidence of systemic involvement on general examination, full blood count, ESR, C reactive protein (CRP) and CT scanning of the abdominal lymph nodes. Duodenal biopsy was also negative. Nonetheless, Whipple’s disease was confirmed on PCR and DNA sequencing of blood and CSF, which are both highly sensitive and specific in our laboratory.

In contrast, patient 2 had secondary Whipple’s disease of the brain also with a complex eye movement abnormality, and myorhythmia, with a raised ESR and CRP, abdominal lymphadenopathy and a lymph node biopsy diagnostic of Whipple’s. This was confirmed on PCR of CSF and blood, and by DNA sequencing. Of interest is that two MR scans, including a 3-Tesla study with gadolinium, showed no abnormalities despite his marked neurological features. The clues to the diagnosis were the raised inflammatory markers and the polyclonal increase in gamma globulins, leading to a search for an underlying lymphoma and the discovery of the intra-abdominal lymphodenopathy.

MRI was useful in the patient with primary Whipple’s of the brain because the nodular enhancing lesions led us to consider Whipple’s disease, but it was unhelpful in the patient with secondary disease because it was normal. The molecular biological tests were essential in establishing the diagnosis in both cases. The lymph node biopsy was helpful in the patient with secondary Whipple’s disease of the brain but there was no tissue confirmation in the patient with primary Whipple’s disease of the brain; it was not possible to do a brain biopsy at the time of her rapid and life-threatening deterioration.


Neurological features can be found in about one third of patients with systemic Whipple’s disease but they are very diverse; cognitive changes, supranuclear ophthalmoplegia and altered levels of consciousness are the most common (table 1).3 A comprehensive analysis of the literature on patients with confirmed primary Whipple’s disease of the brain revealed that the most common manifestations are hemiparesis, cognitive dysfunction, seizures and eye movement abnormalities (table 2). Clearly, on the basis of the clinical features, it is not possible to distinguish primary from secondary Whipple’s disease of the brain because there is so much overlap.

Table 1 Neurological features of systemic Whipple’s disease (% of those of patients who have neurological involvement)
Table 2 Neurological features in 17 patients with confirmed primary Whipple’s disease of the brain (patients may have several features)

Rather, the diagnosis of Whipple’s disease of the brain depends on a low threshold for investigation. PCR for T whipplei on CSF and blood should be considered in the work-up of complex symptomatology, including cognitive change, hemiparesis, seizures and eye movement abnormality which may or may not be associated with oculomasticatory myorhythmia. In the context of such neurological features, MR brain scanning should of course be performed but it may be normal as in patient 2. And even if abnormal the findings are non-specific with multiple nodular enhancing lesions, as found in patient 1, or may be a solitary space occupying lesion.2

The diagnosis of secondary Whipple’s disease of the brain may sometimes be confirmed by biopsy of an abdominal lymph node as in patient 2, and as in both primary and secondary cases, by the application of molecular biological techniques to confirm active infection in the central nervous system (CNS) In the presence of clinical suspicion of Whipple’s disease of the brain, PCR should be performed on the CSF and blood even if the brain MRI is normal. Sometimes brain biopsy may be necessary if the diagnosis is still suspected and the PCR is negative. The sequencing of T whipplei should be attempted on blood and CSF only if the PCR is positive in a laboratory in which these tests have high sensitivity and specificity. Molecular biology techniques are especially important in certain situations, such as in patient 1 where tissue was not available for biopsy. We believe that patients with Whipple’s disease of the brain are likely to be overlooked if this diagnostic approach is not used, and that this disease may be more common than we realise (table 3). The PCR on CSF and blood can also be used to monitor the response to treatment. It is proposed that even with the slightest suspicion of Whipple’s disease of the brain, PCR on CSF and blood should be performed. The test is relatively cheap and relatively easy to perform, and available in most university teaching hospital laboratories.

Table 3 The diagnosis of Whipple’s disease of the brain

Therefore, the diagnosis of Whipple’s disease of the brain depends on:

  • Asking the question: could it be Whipple’s?

  • Looking for evidence of systemic involvement.

  • PCR and DNA sequencing for T whipplei in CSF and blood.


Whipple’s disease of the brain can be fatal if untreated and neurological symptomatology, especially eye movement disorder, may persist after MRI and even molecular evidence of eradication of infection. Unfortunately, the treatment is as difficult as the diagnosis. Therapeutic advice is limited by the lack of randomised controlled trials, unsurprisingly given the small numbers of patients. A generally accepted regime is provided in table 4.

Table 4 Antibiotic guidelines to treat Whipple’s disease of the brain

Retrospective analyses on small series of patients have suggested that cotrimoxazole (trimethorprim-sulfamethoxazole) is one of the best oral treatments, and in tissue culture this is more related to the sulfamethoxazole than the trimethoprim.4 Doxycycline and hydroxycholoroquine are bactericidal against Coxiella burnetii, a bacterium also found in intracellular acidic vacuoles like T whipplei.5 However, doxycycline and other tetracyclines do not cross the blood brain barrier and hydroxychloroquine accumulates in the CNS to a lesser extent than in other organs.6 Although these drugs might be useful in systemic Whipple’s disease without neurological involvement, the reduced penetration of the CNS could be responsible for reported relapses in brain Whipple’s disease from 2% to 33% over five years.7, 8

Schnider et al have shown that third generation cephalosporins are effective in Whipple’s disease of the brain and recommended that ceftriaxone be combined with streptomycin.8 There was lack of response in 40% of patients treated with trimethoprim-sulfamethoxazole. These authors recommended that if neurological relapse occurs during treatment with trimethoprim-sulfamethoxazole, an oral third generation cephalosporin might be useful. Seven of the 15 patients evaluated in this study unfortunately died, confirming the grave nature of Whipple’s disease of the brain.

The European Network on T whipplei infections9 has recently completed the first prospective non-randomised trial comparing parenteral meropenem (3 g daily) with ceftriaxone (2 g daily) for the first two weeks of therapy, followed by maintenance with oral trimethoprim-sulfamethoxazole for the first year in 40 patients (published results not yet available). These antimicrobials penetrate the CNS and are generally desirable in the context of Whipple’s disease of the brain and should be recommended. The required duration of therapy is unknown and patients must be assiduously followed-up.

For patients who are allergic or do not respond to ceftriaxone or meropenem, chloramphenicol or cefepime can be used.10 If meropenem and ceftriaxone are not possible then streptomycin with penicillin G (2 g and 1.2 million units daily respectively) should be considered for two weeks. Interferon-γ has been used successfully in one patient with relapses despite appropriate antibiotic therapy.11

The first isolation of T whipplei from cerebrospinal fluid of two patients in 2003 showed that viable bacteria may survive in the CNS after prolonged antibiotic therapy.12 The two patients did not have neurological symptoms or signs. One specimen was obtained before treatment, the other at relapse 12 months after cessation of therapy. These observations suggest that the organism may be isolated from the CNS in the absence of clinically obvious infection, that all patients with Whipple’s disease should be regarded as having subclinical brain involvement and treated as such, and that patients may have evidence of CNS infection even after prolonged treatment and relapse.12


  • Whipple’s disease of the brain is a rare complex neurological condition with protean manifestations.

  • The distinction between primary and secondary Whipple’s disease of the brain is based on evidence of systemic involvement in the latter.

  • The diagnosis is difficult to make unless it is first considered and then confirmed with tissue biopsy and/or molecular biological techniques.

  • Further studies are necessary to provide an evidence-based foundation to antimicrobial therapy.


I wish to thank Dr Leonie C Reading for permission to show the histopathology. This article was reviewed by Geraint Fuller, Gloucester, UK.


View Abstract

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.