OUP user menu

Primary Whipple's disease of the brain: characterization of the clinical syndrome and molecular diagnosis

P.K. Panegyres, R. Edis, M. Beaman, M. Fallon
DOI: http://dx.doi.org/10.1093/qjmed/hcl081 609-623 First published online: 12 August 2006


Background: Whipple's disease (WD) of the brain without evidence of systemic involvement is a rare illness that is difficult to recognize and potentially life-threatening.

Aim: To elucidate the clinical features and diagnosis of primary WD of the brain.

Design: A single case study, with review of published data.

Methods: We linked the information about our patient with 956 citations to published WD material. We were able to identify 19 other patients with primary WD of the brain.

Results: Our patient was a 48-year-old woman who presented 2 years ago with generalized tonic/clonic seizures. WD of the brain was diagnosed after a life-threatening subacute deterioration leading to reduced consciousness and eye movement abnormalities. She had atrophy and gliosis of the right hippocampal formation, and nodular enhanc-ing lesions. She developed the syndrome of inappropriate ADH secretion, blepharospasm with a complete paralysis of vertical gaze, a severe amnesic syndrome, obstructive sleep apnoea, altered sleep physiology and CSF oligoclonal bands. Primary WD of the brain was diagnosed after PCR confirmed Tropheryma whipplei DNA in CSF and blood. She recovered after intravenous methylprednisolone, meropenem and cotrimoxazole. She has now survived for 24 months, lives independently and drives. Comparing our patient with the 19 others, two clinical syndromes were apparent, in both adults and children: (i) multifarious neurological symptoms and signs with a CT or MRI showing multiple nodular enhancing lesions; (ii) focal neurology secondary to solitary mass lesions.

Discussion: Primary WD of the brain may be diagnosed by recognition of these two clinical syndromes, and confirmed by the application of molecular biological techniques such as PCR.


Whipple's disease (WD) is a rare multisystem infectious disease caused by a soil-borne Gram-positive bacillus Tropheryma whipplei, related to the family Actinomyces.1

WD is usually characterized as a gastrointestinal and rheumatological disorder. Abdominal pain, diarrhoea, weight loss, malabsorption, wasting, low-grade fever, arthralgia, increased skin pigmentation and peripheral lymphadenopathy have all been well recorded since Whipple's original description.2

The nervous system may be involved in around 10–43% of patients with multisystem WD.3–9 Diagnosing central nervous system (CNS) involvement may be difficult in the setting of multisystem WD, and is even more difficult in the very rare cases where there is no evidence of multisystem involvement (primary WD of the CN).8 WD is a diagnosis often proffered to explain mysterious neurological phenomena, but is rarely proven. We consider the clinical and other observations that may help the clinician to consider this diagnosis, in the context of an illustrative case, and review of the literature.

Case history

A 48-year-old mother and day-care supervisor aged 46 years was admitted on 25 November 2003 to Joondalup Health Campus having had two generalized tonic-clonic seizures. Prior to the seizure, her family reported she had progressive problems with short-term memory over several months. Table 1 summarizes the clinical, laboratory and MRI data.

View this table:
Table 1

Cerebral Whipple's disease: laboratory data and MRI findings

The neurological examination was normal, apart from significant short-term memory impairment on bedside examination. She was unable to recall having seen her medical attendant the day before.

MRI revealed an unilateral abnormality within the right temporal lobe involving the right hippocampal formation and amygdala (Figure 1). The amygdala and hippocampus were expanded and oedematous. There was mild uncal expansion and signal hyperintensity. Following gadolinium, there was patchy nodular right temporal lobe cortical enhancement. An EEG revealed spike and wave discharges in the right frontal and temporal region. CSF examination revealed a pressure of 15 cm H2O, with clear colourless fluid, no white cells and 5 red cells per µl. No bacteria were seen. Culture was negative. Protein was 0.41 g/l (0.15–0.45), glucose 2.8 mmol/l (2.8–4.4). Cytology revealed no abnormal cells. On 25 November, her sodium was 126 mmol/l (134–146). Urine osmolality was 711 mmol/kg, serum osmolality 282 mmol/kg (275–295).

Figure 1.

MR images obtained 3 days after presentation. a Transverse T2-weighted image showing right mesial temporal lobe oedema and swelling. b Coronal FLAIR image demonstrating right hippocampal oedema. c,d Transverse and axial contrast-enhanced T1-weighted images showing nodular mesial temporal enhancement.

Around 13 December 2003, she complained of diplopia and ataxia. MRI was repeated on 23 December 2003, showing significant improvement in the oedema and swelling of the right mesial temporal lobe and uncus, attributed to the suppression of seizure activity with anticonvulsants (Figure 2). The right hippocampal body, head and tail revealed further atrophy with asymmetrical dilatation of the temporal horn of the right lateral ventricle. There was mild mass effect involving the uncus with patchy cortical contrast enhancement, and also interval development of cortical signal hyperintensity on T1-weighted images within the right hippocampal formation, consistent with methaemoglobin.

Figure 2.

MR images 4 weeks following presentation. a Transverse T2-weighted image showing persisting right mesial temporal lobe signal abnormality. There is now mild atrophy with dilatation of the temporal horn of the right lateral ventricle. b Coronal FLAIR image showing right persisting hippocampal signal abnormality. c Coronal T-weighted image demonstrates gyral methaemoglobin within the hippocampal head. d Transverse contrast-enhanced T1-weighted images showing persisting nodular mesial temporal enhancement.

Her memory problem continued, with a neuropsychometric profile performed on 29 December 2003 revealing impaired attention, working memory, and speed of information processing with no evidence of executive dysfunction. She showed significant abnormalities in ability to learn and recall new verbal and visual spatial information, incompatible with her previous ability. The pattern was consistent with bilateral mesial temporal lobe involvement.

On 22 January 2004, positive antinuclear factor in a titre on 1:320 with a speckled pattern was noted. Extractable nuclear antigens were positive, with both SSA and SSB detected (anti-Ro and anti-La antibodies respectively). Anti-DNA (double stranded) antibodies were not elevated, at 3 IU/ml. Serum C3 was normal at 1.04 g/l (0.88–1.98) and C4 minimally depressed at 0.15 g/l (0.16–0.52). There was no change in these parameters at the height of her illness in April 2004 (see below). A clinical immunology review did not find evidence of active Sjögren's syndrome or SLE.

The MRI scan on 27 February 2004 showed no change from that of the 23 December 2003.

She was admitted on 4 March 2004 for investigation of the syndrome of inappropriate ADH secretion (SIADH), with a sodium of 125 mmol/l. Her sodium improved with fluid restriction rising to a level of 133 mol/l.

Upon admission, she displayed an amnesic syndrome with impaired episodic memory. She had a complex eye movement disorder, with a moderate eye abduction deficit and impaired saccadic movements in the vertical plain.

She was orthophoric in the primary position, with full vertical eye movements to pursuit. There was saccadic eye movement abnormality on downgaze. There was a slow incomplete eyelid blinks to obtain eye movement. There were full oculocephalic manoeuvres. Convergence was normal. Bell's phenomenon was normal. Pupils were 3 mm on both sides with normal light reaction. The visual fields and fundi were normal. The assessment suggested bilateral supranuclear vertical downgaze saccadic system abnormality, reflective of diencephalic involvement.

On 29 March 2004, she developed blepharospasm, with complete paralysis of upgaze and significantly decreased downgaze not overcome by oculocephalic manoeuvre. The eyes showed poor fixation and were associated with the blepharospasms. She was ataxic, unable to heel-toe walk, with a positive enhanced Romberg's sign. The findings suggested new midbrain diencephalic pathology, and on 31 March 2004, MRI showed right anterior mesial temporal lobe abnormality and atrophic changes as before (Figure 3). There was evidence of nodular signal abnormality extending to involve the right temporal lobe, with extension of the caudate nucleii bilaterally, anterior commissure and right putamen. Abnormalities had also developed in the left insula cortex anteriorly and in the left hippocampal formation. There were nodular and patchy areas of contrast enhancement with gadolinium, involving the right uncus, the middle temporal gyrus and the inferior aspect of the right temporal lobe. Nodular enhancement was also identified in the left anterior insula area and adjacent to the superior margin of the temporal horn of the right lateral ventricular. There was irregular enhancement extending into the perivascular spaces alongside the lenticular striate vessels. Subtle enhancement involving the right insula cortex was identified. There was subtle mesencephalic signal abnormality on T2 and FLAIR images.

Figure 3.

MR images 4 months following presentation. a,b Transverse FLAIR and c coronal FLAIR images showing progressive atrophy of the right mesial temporal lobe with persisting signal abnormality suggesting gliosis. There is now oedema involving the left mesial temporal lobe, caudate heads, and right putamen. d Coronal contrast-enhanced T1-weighted image showing patchy nodular enhancement involving the mesial temporal lobes, insular cortex and extending along lenticulostriate vessels.

There was deterioration in her eye movements after one week, with the development of horizontal diplopia with exotropia and impaired convergence with convergence nystagmus. There was absent upgaze to pursuit and saccades, with absent Bell's phenomenon. Downgaze was absent to saccade and pursuit. Oculocephalic reflexes were absent in the downward position. There was a mild right-sided ptosis.

On 31 March 2004. her confusion, amnesic syndrome and eye movement disorder worsened. She now had a complete palsy of vertical movement and blepharospasm. There was a positive Romberg's sign. A granulomatous disorder was considered, including sarcoid, inflammatory conditions, atypical infections, a paraneoplastic process and Whipple's disease of the brain.

A CSF examination on the 31 March 2004 revealed a pressure of 17 cm H2O. The CSF was clear and colourless. There were 464 leukocytes/µl (90% polymorphs, 10% lymphocytes), and 30 erythrocytes/µl. Cytology showed significant polymorphonuclear leukocytes and abnormal lymphocytes. PAS stain did not show positive cells. No bacteria were identified on the stained preparations. The protein was 0.03 g/l (0.15–0.45); glucose was 1.6 mmol/l (2.8–4.4). Concurrent blood glucose was 3.1 mmol/l. PCR for Tropheryma whipplei was positive (Table 2). The positive T. whipplei PCR in CSF was also confirmed at a second laboratory (Westmead Hospital in New South Wales).

View this table:
Table 2

Molecular biological findings in cerebral Whipple's disease

By 1 April 2004, her clinical condition had deteriorated such that she had several periods of reduced awareness where she had difficulty recognizing her husband and daughter. There were auditory hallucinations. She was rambling, confused and inappropriate, with a fluctuating consciousness and a GCS varying from 14 to 2. She had some mild twitching of the eyelids and the perioral muscles without vergence oscillations. There was no vertical gaze possible up or down. Lateral gaze was relatively preserved. The pupils were 3 mm on both sides, with reduced light reaction. Tone and power in the limbs were normal. Deep tendon reflexes were reduced globally. She was treated intravenously with meropenem 1 g 3 times a day and cotrimoxazole (sulfamethoxazole 400 mg with trimethoprim 80 mg in each ampoule) three ampoules twice per day diluted into 500 ml normal saline for two weeks. Methylprednisolone 500 mg/day for five days was given simultaneously. Her sodium level rose to 139 mmol/l. Fluid restriction continued. She was also given frusemide 20 mg/day, with two salt tablets per day. Her conscious level and eye movement steadily improved, and by 6 April 2004 she could recognize her doctors and attendants. Her conscious level returned to normal and remained stable. A duodenal biopsy showed no pathology. PCR for Whipple's disease on the duodenal biopsy was negative. She maintained her improvement, but remained confused and disorientated. Eventually over several weeks she was able to sit out of bed and started to engage in conversation and walk independently. Her sodium level remained at 138 mmol/l. Her eye movements revealed improvement of the vertical gaze paresis. Vertical saccades were reduced in velocity and were hypometric. Pursuit was intermittent in the vertical plain. A trans-oesophageal echocardiogram was performed which was normal. An abdominal CT scan revealed no evidence of lymphadenopathy.

After intravenous antibiotics she was given oral treatment (Septrin Forte, sulfamethoxazole 800 mg plus trimethoprim 160 mg per tablet; two tablets twice a day). She tolerated this treatment well, her sodium remained normal and she was fit for discharge on 28 April, having only a mild eye movement abnormality and difficulties with episodic memory.

She was followed from 30 April 2004 until 24 December 2004, and developed severe obstructive sleep apnoea with hypoxaemia for which she required CPAP. Overnight polysomnography on 22 July 2004 showed severe insomnia and poor sleep efficiency, frequent waking and marked sleep fragmentation thought directly related to cerebral Whipple's. She was given a trial of CPAP therapy which she did not tolerate; a dental splint was effective.

By May 2005, she was able to drive short distances and live independently with a significant amnesic syndrome, after about 13 months of oral cotrimoxazole. Her vertical eye movement abnormality persisted without tremor or ataxia. Repeat MRI showed complete resolution of modular enhancing abnormalities with residual gliosis and atrophic change in the right hippocampal formation (Figure 4). Repeat CSF examination showed <5 leukocytes and erythrocytes per µl, with normal cytology, protein and glucose. Oligoclonal bands were detected. PCR for T. whipplei was negative in both CSF and blood (Table 2). Her sodium remained normal with salt tablets and fluid reduction. She continues to maintain her improvement.

Figure 4.

MR images obtained 13 months following commencement of cotrimoxazole. a Transverse T2-weighted image and b coronal FLAIR image showing persisting right mesial temporal gliosis and atrophy. c Transverse and d coronal contrast-enhanced T1-weighted images showing resolution of the previously demonstrated nodular contrast enhancement.


Literature search

Examination of the published literature and on-line citations to Whipple's disease identified over 956 references (from PubMed keyword searches of ‘Whipple's disease’ and ‘Whipple's disease of the brain’). We carefully scrutinized the reports that identified involvement of the central nervous system and found 19 patients, or 20 including the case reported above (Table 3). Patients were not included if there was evidence of systemic involvement, e.g. malabsorption, lymphadenopathy, arthritis or positive biopsy findings in the gastrointestinal system.

View this table:
Table 3

Patients with primary Whipple's disease of the central nervous system (n = 20)

The two syndromes of primary WD of the CNS

Analysis of the published data (Table 3) showed that patients with primary WD of the CNS manifest two recognizable syndromes, which may occur in children or adults: (i) multifarious neurological symptoms and signs, with multiple enhancing lesions on CT or MRI (n = 13; 13/18 = 72%); (ii) focal neurology secondary to solitary mass lesions (n = 5, 5/18 = 28%). The total number used for calculations was n = 18, because one of the 20 patients (patient 7, see below) did not have the diagnosis of WD confirmed, and another (patient 11) was originally worked-up in the pre-imaging era.

One patient (Table 3, patient 7) with supranuclear vertical gaze palsy, extrapyramidal syndrome with oculomasticatory myorhythmia had normal MR and CSF.15 However, the diagnosis was not confirmed using tissue pathology or molecular biology, has low probability for the diagnosis of primary WD of the brain, and might represent a neurodegenerative disorder. Unfortunately this patient was lost to follow-up.

Patients with multiple enhancing lesions (type I WD) have a spectrum of clinical features which can include generalized tonic/clonic seizures, ataxia, vertical eye movement disorder, amnesic syndrome, SIADH, obstructive sleep apnoea, insomnia, subacute deterioration, meningoencephalitis, hemiplegia, hypersomnolence, hypogonadotrophic hypogonadism, papilloedema, headaches, vomiting, VII nerve lesions, pyramidal weakness, sensory inattention, enhanced deep tendon reflexes, dementia, weight loss, bizarre affect, dysarthria, nystagmus, Babinski signs, dysmetria, diplopia, absent ankle jerks, reduced sensation glove/stocking distribution, limbic seizures, reduced attention, sudden motor excitement, thirst and fever. The SIADH in our patient is thought to have been secondary to vasopressin release from the neurohypophysis, through cytokine-mediated mechanisms, as found in other forms of encephalitis and meningitis.

Patients with focal mass lesions (type II WD) present with clinical phenomena that depend upon the location of the mass and the findings of raised intracranial pressure; they may include headaches, speech disturbance, impaired concentration, irritability, aggressiveness, partial seizures, hemiparesis, aphasia, Babinski signs, personality change, dysarthria, facial palsy, upper motor neuron signs, behaviour disorder, and unilateral clonic movements.

Life-threatening deterioration

Our patient developed a subacute life-threatening deterioration associated with meningoencephalitis, a CSF pleocytosis with polymorphonuclear leukocytes and a significantly reduced protein, probably as a result of raised intracranial pressure.28 Meningoencephalitis has been observed in three other patients with type I WD (Table 3, patients 4, 11 and 17) and has also been recognized in systemic WD involving the CNS.29,,30 The very low protein of 0.03 g/l found in our patient has not previously been recorded. She was given pulse methylprednisolone to halt the intensity of the inflammatory process and the effects of raised intracranial pressure, as a life-saving measure. The use of corticosteroids in WD of the brain is controversial; however, there might be a role in life-threatening situations.7,,8 We do not advocate the chronic use of steroids.

Our patient's exacerbation demanded that the diagnosis be achieved rapidly without cerebral biopsy. The diagnosis was suspected because of the progressive ophthalmoplegia with MRI changes, and was achieved using PCR of blood and CSF. The majority of recorded patients with primary WD of the brain have been diagnosed by brain biopsy showing characteristic pathology with PAS-positive inclusions, confirmed by ultrastructural studies in some patients. PCR was performed in four other patients (Table 3, patients 2, 4, 10 and 20). In patient 4, PCR in blood and DNA extracted from brain tissue was positive and confirmed by brain pathology. In patients 2, 10 and 20, the PCR on CSF was negative or ‘inconclusive’. In certain clinical situations, tissue examination might not be possible. We believe that PCR on blood and CSF should be seriously considered when attempting the differential diagnosis of primary WD of the brain. This position is supported by the useful application of these techniques to systemic WD when it affects the brain.31,,32 PCR for WD has shown both sensitivity and specificity for confirming a diagnosis of WD.33

Application of PCR to diagnosing primary WD of the CNS

The application of PCR to the diagnostic work-up of patients might help to expand the clinical spectrum of WD of the brain and guide therapy, and there is evidence to suggest that PCR might help in monitoring the patient's infection.31

Molecular diagnosis of WD was first described by Wilson34 using universal bacterial primers, followed by sequencing of the product. This was validated by assays using specific primers for T. whipplei.35 However, T. whipplei PCR has sometimes been positive in patients without WD.36 A related but distinct organism (Whipple's-disease-associated bacterial organism or WABO) has also been associated with WD.37 Coinfection with T. whipplei and WABO has been reported in WD.38 The molecular techniques need to be added to neuroimaging (especially MRI) and CSF examination for diagnosis in those situations where tissue biopsy might not be possible. This is supported by studies of multisystem WD with brain involvement and unhelpful tissue pathology.39 Using this combination of techniques, WD of the brain might become more easily recognizable.

Natural history and treatment of WD of the CNS

The 20 patients were followed-up for periods ranging from 19 days to 23 years. Our patient has survived 24 months. Outcome for patients with type I syndrome with multiple enhancing lesions has been followed from 19 days to 10 years with variable outcomes, including survival and living independently. as well as death. Patients with focal mass lesions (type II) have survived for 3 months to 23 years. The differences in follow-up data might relate to neurological vs. neurosurgical discovery, management and speed of publication, and more long-term data are required.

Reports of the effects of treatment are also fragmentary. No published controlled trials in WD exist, but a comparative study of ceftriaxone and meropenem is currently being conducted by the European Network on T. whipplei infection. Uncontrolled retrospective series of treatment of WD have shown trends for cotrimoxazole being superior to tetracycline40 and penicillin/streptomycin being superior to tetracycline.41 It is recommended that regimens should use antibiotics that cross the blood–brain barrier readily (i.e. penicillin, cotrimoxazole, ceftriaxone, chloramphenicol, quinolones).42 Penicillin/streptomycin or cotrimoxazole have been advocated as initial therapy.43,,44 Experience from WD involving the gastrointestinal tract in which patients have relapsed in the CNS suggests that tetracycline or penicillin should not be used alone.41 It is not known whether newer lipid soluble tetracyclines such as doxycycline would suffice. In vitro sensitivity studies using cell culture have predicted that doxycycline, penicillin and cotrimoxazole have activity (but imipenem and cephalosporins do not).45 Similar studies using axenic medium for culture have confirmed susceptibility of these agents as well as streptomycin and levofloxacin, but cast doubt on cephalosporin activity.46 Our patient was given intravenous cotrimoxazole, meropenem and intravenous methylprednisolone at the height of her illness. She has since been maintained on oral cotrimoxazole with favourable results. Other patients with the type I syndrome have been treated with oral cotrimoxazole (n = 7); penicillin then streptomycin (n = 2); chloramphenicol then tetracycline (n = 2); minocycline with prednisolone, then tetracycline (n = 1). Patients with the type II syndrome with a focal mass lesion have been treated with cotrimoxazole (n = 4) and trimethoprim (n = 2). In the other patients, treatment was not described. Cotrimoxazole failures are also recognized.40

Oligoclonal bands in diagnosing primary WD of the CNS

Our patient developed oligoclonal bands indicative of antibody synthesis within the CNS, as found in multiple sclerosis, cerebral lupus, HIV infection and sarcoidosis; diagnoses excluded in our patient. Oligoclonal banding has also been seen in one other patient with primary WD of the CNS (Table 3, patient 12). We propose that oligoclonal bands may be useful in monitoring the response to treatment, and indicative of chronicity of infection in association with PCR in CSF and blood, although there is little in published studies to guide us in this respect. Withdrawal of treatment has been associated with relapse in the brain in multisystem WD, 2 years after therapy.47

Oculomasticatory myorhythmia in primary WD of the CNS

Oculomasticatory myorhythmia (OMM) is a concurrence of pendular, usually vergence oscillations of the eyes and concurrent contractions of the masticatory muscles. The pendular vergence oscillations are associated with a vertical saccadic palsy.48 OMM has been suggested as a unique involuntary movement disorder involving the extraocular and jaw muscles in WD.49–51 This has been documented in systemic WD, when it involves the brain. One patient (Table 3, patient 11) with primary WD proven by biopsy developed OMM. Interestingly, this patient died in 1931 and was reanalysed in 1993, when the diagnosis was established using pathological criteria. The other patient who developed OMM did not have the diagnosis confirmed and no follow-up information was provided, casting doubt as to the diagnosis of cerebral WD (Table 3, patient 7).

WD involving the brain often causes a deficit in ocular mobility that may mimic those found in progressive supranuclear palsy. Initially, vertical saccades and quick phases are involved, indicating a supranuclear ophthalmoplegia, but eventually all eye movements may be lost. The eye findings in our case strongly suggested the diagnostic possibility of primary WD.

Immune abnormalities for WD of the CNS

Our patient's immune abnormality with a positive speckled antinuclear factor, extractable nuclear antigens, and anti-Ro and anti-La antibodies, might have predisposed her to develop WD. There is evidence for immune abnormalities in WD patients,44 including reduced helper T-cell function,52 delayed-type hypersensitivity reactions and impaired macrophage activity.53 SLE predisposes to many infections, but has never been associated with Whipple's disease. Immune modulation with interferon-gamma has been used successfully in WD.54

Comparison of primary WD of the brain with multisystem WD

The clinical phenomenology of primary WD of the brain shares features with multisystem WD in which supranuclear ophthalmoplegia, memory impairment, confusion and apathy are most common.4,6,,7 That is, the neurological phenomena do not enable a distinction between primary or secondary Whipple's disease of the brain. This distinction must rest upon clinical or laboratory evidence of multiple system involvement.


We have described a patient with primary WD of the brain who presented with generalized tonic/clonic seizures and an evolving neurological syndrome over months with progressive MRI findings and a precipitous deterioration almost resulting in death (Table 1). Her clinical profile has been used as a focus to examine other patients with primary WD of the CNS, in order to develop a syndromic approach to this obscure but treatable illness that may be of use to clinicians. We recommend more widespread use of PCR to search for WD in progressive enigmatic neurological presentations with an appropriate clinical syndrome.


View Abstract