METHODS

Prospective identification of patients presenting to outpatient or emergency departments of a tertiary care centre in Saskatoon, Saskatchewan, Canada for the period of July 1998 to January 2002 was performed. The estimated size of the referral area for this tertiary care centre is 500 000. All patients complained of spontaneous onset of symptoms affecting one side of the body, including head and both limbs. Complaints could include any of numbness, sensory misperception, tingling, heaviness, weakness, or clumsiness. Patients with a history of presenting symptoms suggesting trauma, headache, or seizure were not considered. Patients with a past history of complex regional pain syndrome (CRPS), chronic pain syndrome, or issues with litigation were excluded. Physical examination was performed in all cases, and patients were excluded from the study if the examining neurologist found cognitive changes, motor findings, reflex abnormalities, or dyscoordination. Evidence of cortical sensory loss including extinction on double simultaneous sensory stimulation, agraphesthesia, or astereognosis also led to exclusion from the study. Other modalities of sensory examination performed included comparison of pinprick, light touch using a cotton stick, temperature sensation using a portion of cold metal, and vibratory threshold examination with a 128 Hz tuning fork over both sides of the body, as well as comparison of proprioception between both upper and lower extremities.

All patients were booked for investigations to be performed within 14 days of the onset of symptoms. Investigations performed included computed tomography (CT) and magnetic resonance imaging (MRI) of the brain, lumbar puncture with cerebrospinal fluid (CSF) examined for cell count, protein, cytology, and electrophoresis, carotid ultrasonography, echocardiography, electroencephalography (EEG), somatosensory and visual evoked potentials, and transcranial Doppler ultrasonography with bubble testing to examine for presence of a cardiopulmonary shunt. MRI sequences used included T1 and T2 weighted imaging, diffusion weighted imaging, and proton density imaging. Blood tests including complete blood count, electrolytes, urea, creatinine, serum glucose, creatine kinase, calcium, magnesium, thyroid stimulating hormone level, and vitamin B12 levels were performed in all patients. All patients were suggested to have psychiatric assessment during the study.

After initial investigations, diagnosis, and treatment if necessary, patients received follow up with a repeated history and physical examination after investigations were complete. Long term follow up was conducted between 6 to 18 months after completion of investigations either by a history taken by telephone or by outpatient assessment. Physical examination was performed for those patients seen in outpatient assessments as well as for those who complained of symptoms during the telephone interview.

RESULTS

A total of 34 patients, 25 (74%) women, were identified over the 42 month period. The mean (SD) age of the patients was 35 (11) years. The symptomatic side of the body was the left side in 23 cases (68%). The duration of symptoms prior to assessment was an average of 1.6 (1.2) days.

Initial symptoms consisted of unilateral numbness only in 12 patients (34%), unilateral numbness and tingling in 10 patients (28%), unilateral mixed sensory changes with perception of unilateral body heaviness in nine patients (25%), unilateral sensory changes with perception of unilateral body heaviness and clumsiness in three patients (8%), and unilateral tingling only in two patients (5%). Onset was described as maximal at the start of symptoms in 12 (35%) of patients and as gradually progressive over hours to days in remaining patients (65%). Intermittent blurring of vision in the ipsilateral eye (asthenopia) was described by 10 patients (28%) and ipsilateral disturbance of hearing was described by six patients (16%).

Medical history was remarkable for migraine in six patients (18%), hypertension in six patients (18%), treated hypothyroidism in two patients (6%), hyperlipidaemia in one patient (3%), systemic lupus erythematosus (SLE) in one patient (3%), a prior incidence of undiagnosed hemisensory syndrome in one patient (3%), generalised epilepsy in one patient (3%), and prior pseudoseizures in one patient (3%). Psychiatric history was remarkable for major depression in nine patients (26%), prior personality disorder in six patients (18%) (borderline diagnosis in thee patients (9%), histrionic personality disorder in two patients (6%), and avoidant personality disorder in one patient (3%)), generalised anxiety disorder in two patients (6%), dysthymic disorder in one patient (3%), conversion disorder in one patient (3%), and bipolar affective disorder in one patient (3%).

All patients examined were found to have reduced perception of pinprick and light touch over the symptomatic side. Twenty three patients (68%) identified an increased vibratory threshold relative to the asymptomatic side when vibratory stimulus was placed at the digits, sternum, and frontal bone. Three patients (9%) showed loss of proprioception at digits of the upper and lower extremities on the symptomatic side as compared with the asymptomatic side. No patients identified a loss of temperature perception over the symptomatic side. Comparison of pinprick over the torso showed an abrupt border between abnormal pinprick sensation on the symptomatic side and normal pinprick sensation over the asymptomatic portions in 27 (79%) of patients.

All 34 patients had blood tests, CT imaging of the brain, MRI of the brain, and carotid ultrasonography. Additional diagnostic procedures were performed in less than 100% of patients and are shown in table 1. Both the neurologist and patient knew any abnormalities of investigations before follow up assessment in all cases. Abnormalities found within blood tests in five patients (15%) were deemed non-contributory. Further testing was performed in one patient (3%) with a diagnostic CT scan finding and revealed a positive antiphospholipid antibody profile.

CT of the brain was abnormal in four patients (12%) and in one case, the abnormal finding was deemed significant—a single abnormal hypodensity in the corona radiata contralateral to the hemisensory loss. This patient had a history of SLE and presented with unilateral hemibody numbness only. MRI revealed abnormalities in the same four patients found to have CT scan abnormalities, with again one patient having an abnormality deemed to be significant—a single oval T2 and proton density hyperintensity in the corona radiata contralateral to the hemibody sensory disturbance, in the same location as was noted on this patient’s CT scan. Diffusion weighted MRI similarly demonstrated a hyperintense lesion in this patient.

After completion of investigations, all patients had a follow up history and physical examination performed an average of 9.6 (5.8) days after presentation. Of the 34 patients, 16 patients (47%) continued to complain of hemisensory difficulties. Physical examination at that time showed reduced perception of pinprick and light touch in all 16 patients, with raised vibratory threshold on the symptomatic side relative to asymptomatic side in 11 patients (32% of all patients), all of whom had reported this deficit at presentation. The remainder of examination was normal in all patients at this initial follow up period.

Long term follow up by telephone was available for 30 (88%) of 34 patients at 16 (7) months. Persisting symptoms were identified in 6 of 30 patients (20%), including unilateral numbness in three patients (10%) and unilateral tingling in three patients (10%). One of the six patients had a second clinical syndrome consisting of right hemiparesis diagnosed as stroke that occurred seven months after first presentation. This patient was the same person identified with positive neuroimaging previously and a history of SLE. Since initial assessment, the patient had been found to have positive IgG lupus anticoagulant antibodies with a titre of 1:400, and was subsequently treated with anticoagulation after the occurrence of stroke with no further events over the following 13 months. No other patients developed any discrete neurological events during long term follow up. Examination was performed in a total of 10 patients, including all six patients that had complained of persisting symptoms. Examinations in asymptomatic patients were normal in all four cases. Two of the six (33%) symptomatic patients seen at 12 (5) months after initial assessment had hemisensory pinprick and light touch loss relative to their asymptomatic side, with the remainder of their examination normal. The two patients with abnormal examinations at this time continued to have a very sharp transition between abnormal and normal regions of pinprick sensation.

Physicians other than the author made subsequent diagnoses of migraine without aura (4), migraine with aura (2), major depression (2), generalised anxiety disorder (1), and ipsilateral carpal tunnel syndrome (1) in these patients over the course of follow up. Twelve (35%) of 34 patients agreed to psychiatric assessments after investigation completion. Of those 12 patients agreeing to psychiatric assessment, six (50%) had no psychiatric diagnosis. In the remaining six patients, two were diagnosed with histrionic personality disorder, two with borderline personality disorder, one with major depression, and one with dysthymic disorder.

DISCUSSION

I report my experience with 34 patients who presented with hemisensory syndrome during a period of 42 consecutive months within a referral area having a population of 500 000. Thus, the estimated incidence of hemisensory syndrome in our geographical area is two patients per 100 000 per year. This estimate is probably low, as not all patients with hemisensory syndrome would have necessarily been referred to this tertiary care institution during this time period.

An abrupt change in sensation, or exact splitting at the midline, was described by most (79%) patients. Although organic disease has been reported as a cause of this pattern,^5,6 cutaneous branches of the intercostal nerves typically overlap 1 cm–2 cm, and sensory loss from an organic lesion would be expected to be paramedian.⁷ Most patients (68%) in this study also demonstrated a difference in vibratory threshold between affected hemibody and unaffected side. Although commonly used by neurologists, examination using vibration over the left and right portions of a single midline bony structure such as the sternum or frontal bone is often described as asymmetric by patients with organic disease.^6,8

Hemisensory deficits as presented in this study, can occur after pathological insults to the contralateral thalamus or parietal cortex. The thalamus, in particular, receives sensory input from the entire contralateral body, and lesions of the ventral posterior nuclear groups of the thalamus are associated with a contralateral hemisensory deficit of tactile, temperature, and pain sensation.⁹ Pure hemisensory deficit has been described with stroke of both ischaemic^10,11 and haemorrhagic¹² aetiology. Neuroanatomical localisations that could produce a pure hemisensory deficit include the ventroposterior thalamus,^13,14 corona radiata,⁹ paramedian dorsalateral pons,¹⁰ lenticulocapsular region,^9,12 and parietal lobe.¹¹ Yet, aetiologies such as these were only found in one patient in this study. Further specialised diagnostic procedures may contribute to diagnosis in future patients, such as functional MRI (A Mailis, et al, American Pain Society, 2000), which has not demonstrated the expected activation of the contralateral somatosensory cortex with somatosensory stimulation of symptomatic limbs in patients with chronic pain syndrome and somatosensory deficit. Perhaps Charcot was somewhat correct when he believed that hysterical phenomena affecting a hemibody indeed has a dysfunction “in the grey matter of the cerebral hemisphere on the side opposite”.³

Although organic neurological diagnoses could explain the existence of a hemisensory syndrome, only one patient (3%) was diagnosed with a specific neurological diagnosis after substantial diagnostic procedures and follow up periods. Thus, the yield of diagnostic testing in hemisensory syndrome is very low. The explanation for symptoms occurring in remaining patients is unclear. The role of psychological processes such as anxiety, depression, or stress may influence the central nervous system and its perception of sensory stimuli, as has been suggested with increased nociception in CRPS (A Mailis, et al, American Pain Society, 16th annual meeting, New Orleans, 1997). How the role of psychobiological interactions between personality and psychological profiles affects sensory processing functions for one hemiportion of the body only is unclear. Similar clinical presentations have been described in patients with CRPS,^8,15 chronic pain syndromes (A Mailis, et al, American Pain Society, 16th annual meeting), and fibromyalgia (H H S Kaziyama, et al, IASP 9th world congress in pain, Vienna, Austria, 1999). A high incidence of psychiatric illness was diagnosed in patients with hemisensory impairment and CRPS.^15,16 In one study, 84% of patients with hemisensory deficit and CRPS had a psychiatric diagnosis including mood disorder in 46%, anxiety disorders in 30%, and substance misuse disorders in 14%.¹⁶

In this prospective series, the diagnostic yield for specific aetiology of spontaneous hemisensory syndrome was 3% and all other remaining patients had a benign prognosis. I advocate strong consideration of background history, neuroimaging, and reinforcement of the nearly uniform benign prognosis for patients with spontaneous hemisensory syndrome. I would not advocate aggressive investigations beyond neuroimaging in this group of patients unless specific historical points indicate further interventions. Psychiatric evaluation may be beneficial in some cases. The cause of hemisensory syndrome remains obscure and may relate to a combination of central processing difficulties or psychological factors.