Clinical manifestations

Presentation and course

HTLV-1-associated myelopathy/tropical spastic paraparesis (HAM/TSP) is a slowly progressive inflammatory myelopathy that evolves slowly over decades. Although onset is generally during adulthood, age of onset is highly variable and may be as early as the teen years (69) and as late as the late 50s. HAM/TSP affects females more often than males (60/40). Typical features include spastic paraparesis remarkable for early bladder dysfunction, neuritic pains in the low back and legs, and frequent subtle lower motor neuron findings (22).

Symptoms start insidiously with neuritic pains (painful burning sensations) occurring symmetrically in the legs in an ascending pattern. Even more characteristic is the same type of dysesthesia in the lower back, occasionally in a band-like fashion, with sciatica that may have led to back surgery before diagnosis. Bladder dysfunction begins early and is manifested by chronic urinary retention. The patient frequently neglects this symptom because it is most often painless. Nocturia is the most frequent symptom, followed by incontinence and urinary urgency and frequency. Urodynamic testing is abnormal in 80% of patients; the major abnormalities include detrusor overreactivity followed by detrusor-external sphincter dyssynergia diagnosed in 25% of affected individuals. Some carriers had an abnormal cystomanometry, bringing up the possibility that HAM/TSP may be pauci-symptomatic in its early, preclinical phase. Recurrent urinary tract infections may represent the onset of symptoms. We suspect that some patients have died of urosepsis before the diagnosis was ever contemplated. Prospective follow-up of positive blood donors reveals that HTLV-1 positive nonaffected carriers were even more prone than controls to report symptoms such as leg weakness, impaired tandem gait, Babinski sign, impaired vibration sense, or urinary incontinence. The International Research Group has described and validated a bladder dysfunction symptom score for HAM/TSP that should be useful in the follow-up of clinical trials (78).

Over a matter of months to years, spasticity sets in and is often described early on as a slight sense of stiffness. Spasticity slowly evolves towards pronounced extensor spasms and scissoring. In the first few years, spasticity is intense and contrasts with the subtle muscle weakness. Spasticity disappears in the end stage.

On examination, spasticity predominates and is bilateral and symmetric. Muscle weakness affects the legs but spares the arms for a long time. Weakness is generally distal, affecting leg flexors, but it can also be proximal and then leads to suspect an associated HTLV-1 polymyositis or neuritis. Deep tendon reflexes are hyperactive with up-going toes and extensor spasms. Minor proprioceptive deficit is limited to a slight reduction of vibration sense in the lower limbs. Examination of the abdomen may reveal a distended bladder, which is the hallmark of advanced disease. Postvoid catheterization can show increased residual urine, often in the order of 500 to 600 ml (N < 50 ml). The voiding dysfunction often requires self-catheterization.

Lesions of cellular structures in the gray matter of the spinal cord are associated with signs of peripheral nervous system involvement that are frequent in this myelopathy. The leg muscles can show markedly reduced bulk. Some patients can present with an amyotrophic lateral sclerosis-like syndrome (42). Pathology is clearly different, however, from that of amyotrophic lateral sclerosis (03). Sensory discrimination is reduced in a stocking pattern. This is often accompanied by painful dysesthesia with burning sensations on skin stimulation. When progressive and in conjunction with the effects of advancing age, ankle or even knee jerks are lost. One may occasionally find evidence of lesions above the foramen magnum with brisk jaw jerk, nystagmus, and pallor of the optic discs. Interestingly, Brazilian authors have demonstrated that vestibular-evoked myogenic potentials reveal subclinical abnormalities in most patients (33).

Progression of the deficits leads to inability to walk unaided within a median of around 10 to 13 years and an average of 20 years to reach wheelchair-bound status. Twenty-five percent of Olindo's patients were deceased 20 years after the onset of the disease (49). The course is considered rapid if the patient reaches wheelchair status in less than 2 years. The group in Kagoshima has reported on factors that render the disease subacute; this includes onset of the disease at a later age, more inflammatory CSF, and having received large amounts of viruses essentially from transfusion (41). These rapidly evolving cases should be differentiated from neuromyelitis optica and multiple sclerosis.

Death is generally due to septic shock from pyelonephritis secondary to bladder dysfunction or to infected decubiti.

Associated conditions

The carrier state. It is difficult to evaluate how many individuals are carriers. In Japan, it has been extrapolated from seropositivity of blood donors that over 1 million individuals are carriers. This number is now much reduced because vertical transmission has been eliminated by screening out women who are positive, so they no longer breastfeed and do not transmit the virus to their progeny (66).

A non-negligible proportion of patients with HAM/TSP die of adult T-cell lymphoma leukemia (ATLL), another HTLV-1-associated disorder. In a study, ATLL was one of the major causes of death in 5 of 33 patients (45). Having dominant clonal expansion of HTLV-1-infected cells with ATLL-associated somatic mutations may be important characteristics of patients with HAM/TSP who are at risk of developing ATLL. With erythroderma and pruritus associated with lymphadenopathies and possibly leukemia, the prognosis of slowly progressive HTLV-1-associated myelopathy is suddenly transformed into a short course towards demise.

Many conditions have been recognized as being more or less closely linked to HTLV-1-associated myelopathy and the “carrier” state. In regard to the tightness of this linkage, conditions can be divided into those clearly associated with HTLV-1 (either in patients with HTLV-1-associated myelopathy or in carriers), and those in whom the link has not been demonstrated beyond doubt. Clearly associated conditions are polymyositis, interstitial pneumonitis, dermatitis (either staph or generalized scabies), and uveitis. Most probably associated are autoimmune conditions such as peripheral neuropathies, Sjögren syndrome, arthritis, arteritis, and an anticardiolipid syndrome.

Polymyositis. Polymyositis is directly due to the HTLV-1 virus and is associated with infiltration of muscle by CD8+ cells. Generally clinically silent, polymyositis is recognized by increased creatine phosphokinase and is almost always present on biopsy or autopsy. It is generally insidious and latent clinically (44). Essentially, it is present among patients with HAM/TSP but not in carriers. It is thought to be due to an attack by specific natural killer cells on muscle fibers invaded by the virus.

Interstitial pneumonitis. Interstitial pneumonitis goes frequently unrecognized; it affects carriers and patients with HAM/TSP equally. It is marked by a chronic irritative cough, possibly from bronchospasm. Bronchoscopy and bronchial lavage suggest the diagnosis if there is a predominance of alveolar macrophages. Severity of this disease parallels provirus load (80).

Dermatitis. Staphylococcic dermatitis was originally described by La Grenade (34). It is a chronic dermatitis affecting young carriers during childhood. It is a generalized rash with exudates and crusting on ears, eyelids, neck, axilla, and groin. Staphylococcic dermatitis is accompanied by a watery nasal discharge and lymphadenopathy. Cultures are frequently positive for beta-hemolytic streptococcus or staphylococcus aureus. Dermatitis should be differentiated from eczema and atopic dermatitis. It responds well to antibiotics and steroid creams but tends to recur. It is more frequently found in children born from carrier mothers (vertical transmission). It is probably secondary to a very focused deficit in immune response. Generalized scabies is an associated opportunistic infection (09) as well as digestive strongyloidiasis, which is occasionally hidden behind vitamin B12 deficiency.

Uveitis. On systematic review, uveitis affects 14.5% of infected patients but has generally little effect on visual function. Uveitis usually affects the anterior or intermediary segments. Keratoconjunctivitis and keratitis can be seen and are often part of the sicca syndrome.

HTLV-1 encephalitis. This term has been coined to describe a single patient with known HTLV-1-associated myelopathy who developed subacute arm weakness, sensory loss, and cerebellar dysfunction accompanied by multiple cortical and subcortical T2-hyperintense MRI lesions (29). She improved with corticosteroids.

Peripheral neuropathy. A peripheral neuropathy that can be found in carriers and in patients with HAM/TSP has been reported by the French and the Brazilian group. It may be multifactorial, as demyelination, inflammatory infiltrates, and axonal degeneration have been found. The French group in Martinique has extended the notion of peripheral nervous system involvement and demonstrated the presence of clinical, electrophysiological, and biopsy abnormalities in most patients (64).

Amyotrophic lateral sclerosis-like syndrome. An HTLV-1-associated amyotrophic lateral sclerosis-like syndrome is rarely seen. In our experience this amyotrophic lateral sclerosis syndrome is rare and associated with a highly inflammatory CSF (03).

Autoimmune disorders. There is little doubt that most of the associated diseases mentioned above have an autoimmune basis and are either due to direct viral infection or an overshoot of the immune response to the virus. An association with nonneurologic autoimmune disorders such as Graves disease, Sjögren syndrome, diabetes mellitus, and autoantibodies has been reported in Jamaica (72). These associations are probably not coincidental and are important on a theoretical as well as practical basis. They strongly support the virally induced theory of autoimmune disorders. This class of disorders may be a direct consequence of CD4 T-cell invasion by the virus, generating an autocrine process through IL-2/IL-2 receptor and IL-12/IL-12 receptor (47). These associations are found both in carriers and in HTLV-1-associated myelopathy patients. Other difficult to ascertain associations include Bell palsy, anticardiolipid syndrome, and xerosis cutis.

Prognosis and complications

HTLV-1-associated myelopathy is a relentlessly progressive disorder leading to triple flexion (or extension) paralysis, decubiti, urinary sepsis, and death in 2 to 3 decades . If the chronic bladder retention is not detected and is left untreated in the context of substandard health care conditions, it generally leads to urosepsis. This, by itself, can lead to rapid demise, which is unfortunately not infrequent in the Third World, where many of these patients live and where the virus is endemic. Patients with HAM/TSP are also at risk to develop adult T-cell leukemia/lymphoma (45).

Sato and colleagues found that CSF cell count, neopterin concentration, and CSF levels of C-X-C motif chemokine 10 correlated well with disease progression over four years, better even than HTLV-1 proviral load in peripheral blood mononuclear cells (PBMCs), but a prospective study is needed to confirm this observation (67).

Clinical vignette

In 1995 a 48-year-old woman was originally seen for leg weakness. She stated that she had been well until 1991, when she started to develop low back pain that referred to the left leg on the posterior and lateral aspects and went down to the foot. Pains and cramps occurred in the leg a few times each week. In parallel, she stated that she was developing weakness that had worsened since 1994. She felt unsteady on her feet and had occasional episodes of urinary incontinence. She had had some urgency and incontinence since the delivery of twins in 1983, but over the last six months she had complete incontinence with total loss of bladder control. She reported no symptoms above her waist. Past medical history was negative for diabetes, hypertension, or allergies. Family history was unremarkable. She was half North Pacific Coast First Nation and half Caucasian.

On examination, she was obese (120 kg for 158 cm) and vitals were stable. Mental status was normal. Visual fields were full with normal fundi. Extraocular movements were full. The rest of the cranial nerve exam was normal. Motor examination was normal in the upper extremities with some hyperreflexia, no sensory deficit, and good coordination. In the lower extremities there was weakness of hip flexion (4/5), knee flexion (4/5), and foot dorsal extension. This deficit was symmetrical and was accompanied by increased tone, increased deep tendon reflexes, and bilateral up-going toes. She had an indistinct sensory level that faded at mid-chest. She had mild decrease of vibration with normal position sense, and she needed two assistants to walk.

Her complete blood count, white blood cells, and differential count were normal; blood smear showed a few "flower cells.” Visual- and brainstem-evoked potentials were normal. MRI of the brain showed a few non-specific areas of increased T2. MRI of the cord revealed borderline atrophy of the thoracic cord (diameter 6.5 mm).

CSF examination showed three white cells/cubic mm and an increased CSF IgG index with oligoclonal bands in CSF and not in serum. Her bladder was atonic with a residual volume of 450 ml. The patient was started on a self-catheterization program. She had antibodies to HTLV-1 by ELISA and western blot both in serum and CSF. Pro-viral DNA was amplified from blood cells and CSF using PCR. She was negative for hepatitis C. She had never received blood products or used intravenous drugs. She had been breastfed by her mother, who was herself HTLV-1 negative. Her contamination probably occurred through sexual activity. She was treated with monthly intravenous methylprednisolone sodium succinate infusions and zidovudine (400 mg orally, daily).

Over the following years, the disease continued to slowly evolve. Some complications arose such as generalized scabies treated with ivermectin tablets. She also developed a single large lymphadenopathy in her neck that was removed and revealed nonspecific inflammation on histological exam.

In 2010, she was wheelchair-bound and had no movements in her legs; she continued to lose strength slowly. Surprisingly there was a major loss of muscle bulk in the legs without increased creatine phosphokinase. It should be stressed that the motor deficit was not ascending but, if she still had normal strength in her upper extremities, she was now developing spasticity in both arms with occasional spasms. These responded well to high doses of antispastic medications.

Biological basis

Etiology and pathogenesis

Pathology of HTLV-1-associated myelopathy/tropical spastic paraparesis includes inflammatory demyelination focused on the thoracic spinal cord but leaching up to the brain. This has been well described by Japanese (52), French (21), and American authors (27). The brunt of the disease appears to fall on the thoracic spinal cord, which eventually becomes atrophic. On microscopic examination, there is an inflammatory infiltrate affecting both grey and white matter with demyelination. The intensity of the infiltrate varies with the duration of disease and is reduced as time passes. Why is the disease principally affecting the middle thoracic cord? There is no good explanation for this; however, this is the area where blood flow is slower, and it is hypothesized that cell extravasation from the blood stream to the spinal cord is increased in this area of low blood flow. This is the same area affected by thoracic lesions that cause the sensory “multiple sclerosis hug.” In the same vein, Taniguchi and colleagues have shown that when cerebral lesions exist, they appear in the watershed area of the brain where blood flow is less (74). The pathogenesis of rare cases of HTLV-1 encephalitis can probably be attributed to a similar mechanism (29).

Amplifying this, lymphoid cells infected with HTLV-1 are activated and exhibit more adhesion molecules than cells derived from healthy controls and patients with multiple sclerosis (02).

The mechanism of tissue destruction is not settled, but most probably involves immune-mediated mechanisms. Certain facts are clearly established: HTLV-1 preferentially infects CD4+ T cells (T helper), and through the TAX gene, HTLV-1 generates an autocrine situation where the infected cells secrete IL-2 and express IL-2 receptors on the infected lymphocytes' membrane. This autocrine activation results in spontaneous DNA synthesis. When blood lymphocytes are put in culture, a unique phenomenon is observed akin to what is seen in some leukemias--the presence of spontaneous DNA synthesis in circulating lymphocytes in the absence of leukemia (40).

Increased spontaneous DNA synthesis by lymphocytes in culture starts at 48 hours and lasts for the six days in culture. This corresponds in vitro to the presence of multinucleated lymphocytes, which, observed under the microscope, are nicknamed "flower cells” (lymphocytes with a burgeoning nucleus). This is so characteristic that the presence of flower cells can be used by hematologists to diagnose this condition. However, although flower cells can be recognized on slides, automatic white blood cell counting machines will generally read them as polymorpho-nucleated cells and not report them as being abnormal. Flower cells are found in carriers and in adult T-cell lymphoma/leukemia patients. Practical interest has started to focus in measuring this phenomenon as a way of measuring in vitro and is used as a marker of the effects of drugs in clinical trials.

In patients with HAM/TSP and carriers, lymphocyte subpopulations differ from healthy controls. The proportion of CD3, CD4, and CD8 is normal. However, there is evidence of cell activation as well as changes in markers of cell function (02). Higher proportion of IFN-γ+ Gamma-Delta T lymphocytes was found in HTLV-1-infected patients, which positively correlated with the HTLV-1 proviral load in peripheral blood mononuclear cells (12).

A second contributor to tissue destruction is akin to “collateral damage,” as there is a very strong killer cell response directed against some of the viral antigens. Viral antigens multiply in T cells and elicit a cytotoxic killer cell response. Of special pathogenic relevance are the changes in cytotoxic T cells or natural killer cells (NKT cells). Jacobson and colleagues have shown that in HAM/TSP, cytotoxic T cells are specific for the HTLV-1 tax antigen (31; 32). On the other hand, it appears that NKT cells are reduced in number in the peripheral blood of individuals with HAM/TSP; fewer are CD8+, and more are CD4+ (46). A unifying theory suggests that HTLV-1 specific killer cells are generated in reaction to the infected CD4 lymphocytes (and other target cells such as myocytes), representing an attempt to limit the infection.

Natural killer cell overactivity is clearly at the origin of the myositis. It is practically always present at some point during the disease. It is not evident that the virus infects cells in the CNS. Indeed, viral particles can be found inside macrophages and T cells, but they have not been found in neurons and glial cells.

Leal and colleagues demonstrated that CD39 is upregulated on CD4+ T cells in HAM/TSP patients, probably playing a role in the development of the proinflammatory milieu (36). Accordingly, they suggested that CD39 upregulation may serve as a surrogate diagnostic marker of progression and could potentially be a target for interventions to reduce the development of HAM/TSP. Measuring plasma inflammatory cytokines may also help monitoring HTLV-1 carriers (13).

The relative contribution of the genetic make-up of individuals and of the structure of the virus in becoming affected with HAM/TSP versus remaining a healthy carrier is still unclear (06; 75).

HTLV-1 is a retrovirus, and it is worth noting that approximately 8% of the human genome consists of endogenous retroviral sequences that have been integrated over the millennia. These sequences have been amplified (35), and some are clearly overrepresented in relapsing multiple sclerosis (15), but such amplification data should not be overinterpreted. There is clear evidence that HTLV-1 is not the cause of multiple sclerosis (60).

Epidemiology

The transmission of HTLV-1 bears similarities to that of HIV, specifically because of vertical and horizontal transmission; however, contrary to HIV, the HTLV-1 virus cannot survive without being cell-associated. It is also very stable over time, ie, it is only rarely subject to mutations, which is a detail of interest in studies of population genetics and thus human migrations. This explains many of the epidemiological features of HAM/TSP.

HAM/TSP predominates in HTLV-1 endemic areas such as southern Japan, the Caribbean, equatorial Africa, and the Seychelles. Other high prevalence foci include Central and South America, Iran, Melanesia, southern Africa, and the Pacific coast of Canada. In North America, endemically infected individuals are either Pacific Northwest natives (48) or descendants of West Indies migrants (56; 71). In these areas, the disease is often associated with a specific ethnic group living in relative isolation and is rarely found outside of this group. Pacific Northwest coastal natives have been isolated for many centuries, akin to people who live on small islands. Although the disease is endemic in limited subpopulations, a number of cases are now being recognized in Caucasians living in the same areas. Whether these individuals are contaminated by blood transfusion, use of intravenous drugs, or sexual activity has not been established. Some groups living in relative isolation, such as Iranian Jews living in Mashhad, are also contaminated with a high incidence (01), probably due to poor hygiene and endogamic sexual activity. However, there seems to be some background of seropositivity where isolated populations can show high seropositivity. This is the case in Mashhad (65) as well as other cities on the Silk Trade Road where non-Jews have a low endemic background seropositivity.

HTLV-1 is transmitted by cells in bodily secretions and blood. It is transmitted vertically from mother to child, probably by breastfeeding rather than by contamination in the birth canal. Adult T-cell leukemia/lymphoma and infectious dermatitis only occur following mother to child transmission, whereas 30% of HAM/TSP arises only from maternal transmission, the rest from sexual contact and blood products (62). With sexual contact, women are more prone to infection than men. In Japanese couples, only 10% of men will become positive after 10 years of common life with an infected woman, whereas 80% of women will become positive if it is the male partner who is positive. Interestingly, among children born from infected mothers, females are more frequently positive and sooner (11).

Transmission through blood transfusion occurs, but as the virus does not survive if not cell-associated, only fresh blood transmits the virus. HTLV-1 virus can also be transmitted by puncture with a blood-contaminated needle. In intravenous drug users, HTLV-1 positivity is often associated with hepatitis C infection. In Canada, blood and organ donors have been screened for HTLV-1 serology since the early 1990s. Similar to HIV1, transmission of HTLV-1 has been reported from wild African monkey bites to African hunters (18).

One of the major biological differences between HIV and HTLV-1 (both retroviruses) is the fact that HTLV-1 mutates much less, and this is to such an extent that it can be used in anthropology to evaluate human genealogies and migration patterns. An example of this is the finding that Canadian Inuits carry a virus closely related to two strains isolated from Eastern Siberian tribes with a separation date going back from approximately 3000 to 7000 years before our time (04).

Three types of HTLV-1 viruses are recognized based on envelope sequencing:

Most interestingly, only a small proportion of the individuals carrying the virus will develop HTLV-1-associated myelopathy. The rate at which carriers can express the neurologic disease was originally considered to be 1 in 400, but this is being challenged: first, by the data from Mashad (Iran), when an elderly population of Iranians of Jewish descent from Mashad was repatriated to Israel. Most were found to be seropositive, and a very high proportion of the carriers exhibited evidence of spasticity. This suggests that aging favors the expression of the disease (01).

Further, an ongoing prospective study of blood donors in San Francisco confirms this trend and reveals that the incidence of HAM/TSP developing among healthy carriers is higher than anticipated (6 cases among 160 HTLV-1 over 10 years) (51). Similarly, a long-term prospective study on a cohort of initially asymptomatic individuals in Brazil found that the incidence of HTLV-1-associated myelopathy was 5.3 cases per 1000 HTLV-1-seropositive cases per year (95% CI: 2.6-10.9) (61). In 2015, the Brazilian group has just released results of their cohort study where they found that 76 out of 414 HTLV-1 positive patients had HTLV-1-associated myelopathy. Among the 251 non-HTLV-1-associated myelopathy patients followed, five developed HTLV-1-associated myelopathy, 187 had feet numbness, 130 nocturia, 127 urgency of urination, 76 leg hyperreflexia, 53 leg weakness, and 37 Babinski signs. Females and individuals with QPCR showing more than 50,000 copies/1,000,000 PBMNC had higher risk of progression towards HTLV-1-associated myelopathy (73).

Silent dissemination of HTLV-1 inside families is well established (19) and can be misinterpreted as evidence of a genetic disorder. The finding that children with early onset are essentially born to HTLV-1-positive mothers argues strongly that early onset is due to vertical transmission (69). Avoiding breastfeeding in HTLV-1-positive pregnant women has totally controlled the vertical transmission of the virus in Japan.

Prevention

Prevention of transmission is done by classical means well-publicized since the HIV epidemics: safe sex (ie, condoms or abstinence from intercourse) and avoidance of needle sharing. For health practitioners, the usual good practices should be applied in safe handling of sharp instruments. Blood and organ donors should be screened for HTLV-1 antibody positivity. This generally has been done in North America and Europe since the late 1980s (28).

Women who are carriers should not breastfeed. Since this recommendation was applied in Japan, cases of HTLV-1 positivity among Japanese children have been drastically reduced. This public health policy of antenatal testing and screening has been shown to be a cost effective method and is known to prevent the full-blown disease from appearing in carriers, but development is said to be rare (1 in 400) (30). This statement may, however, have to be revisited in view of the Mashad story (vide supra) whereby 100% of HTLV-1-positive elderly people in a long-term care facility were found to have spasticity but not full-blown HTLV-1-associated myelopathy. This could indicate that carriers in endemic areas such as Brazil’s favelas and other third world populations may have a reduced life expectancy that limits the ability of the full HTLV-1-associated myelopathy picture to express itself with aging.

Differential diagnosis

Confusing conditions

HTLV-1-associated myelopathy is fundamentally of a progressive spastic inflammatory paraparesis, and attention should be focused on eliminating curable causes, including spinal cord compression, before doing anything else. An attentive search for a sensory level of a segmental abnormality is key. Increased protein without cellularity of the CSF should be a red flag, but imaging is the key to diagnosis. MRI is the test of choice unless there is a suspected level where CAT scan will generally be appropriate.

Amyotrophic lateral sclerosis can be eliminated on the absence of lower motor neuron involvement, either clinically or by EMG. Lumbar puncture is normal in amyotrophic lateral sclerosis.

Early acquired and inherited conditions, such as Strumpell-Lorrain, can be tricky to eliminate unless the disorder starts early, as the onset of HTLV-1-associated myelopathy is practically not seen before puberty (69). The vertical transmission of this disease can be misinterpreted as a familial condition. The lumbar puncture findings of an inflammatory component should rectify the search for an etiology.

The absence of past neurologic complaints helps eliminate multiple sclerosis; however, primary progressive multiple sclerosis can look very similar with its paucity of symptoms and its slow progression. Searching for an old episode of optic neuritis or a remitting sensory event will be the key. The diagnosis is made even more complex because of the presence of a few lymphocytes and of oligoclonal IgG bands. Multiple sclerosis is suspected in the event of abnormal white matter on brain MRI.

Only PCR looking for HTLV-1 will elicit the appropriate final diagnosis.

The clinical experience of this author is that the disease is overdiagnosed in endemic areas but is underdiagnosed in nonendemic areas. In endemic areas, the means of diagnosis are rarely available (ie, MRI of the spinal cord, quantitative PCR), whereas in nonendemic areas the disease is easily missed if the patient does not belong to a population at risk. In developed countries, one should systematically think about this diagnosis. Regardless of ethnic background, the diagnosis rests on serology comparing serum and CSF and on quantitative PCR. These data are difficult to obtain in less developed countries. Unfortunately, epidemiological data are too often based on just the association of spasticity of gait and positive HTLV-1 serology. Clinically, this association may be misleading and represent just chance co-occurrence.

Proviral load is often higher by 2 to 3 logs in patients with HTLV-1-associated myelopathy compared to in carriers. Unfortunately, there is significant variability in proviral loads in each individual from either group. Thus, Furtado and colleagues used receiver operating characteristic (ROC) curve analysis to determine the best cutoff point of the proviral load to differentiate between asymptomatic carriers and patients with HTLV-1-associated myelopathy enrolled in a long prospective study (20). A cutoff value of 11,400 copies/1 million cells (1.14%) was 78.2% sensitive in identifying patients with HTLV-1-associated myelopathy versus carriers.

The diagnosis rests on serology (57; 58). The presence of antibodies to the HTLV-1 and HTLV-2 viruses is evaluated using an ELISA.

In areas where HTLV-1 is not endemic, the likelihood of exposure and the finding of a positive serology should be sufficient for diagnosis in the presence of an inflammatory myelopathy. The diagnosis of inflammatory myelopathy is made when there is a progressive myelopathy with CSF lymphocytosis, increased synthesis of CSF IgG and oligoclonal bands, and if the MRI rules out anatomic causes. The diagnosis of HAM/TSP is then confirmed by the positive serology.

This eliminates progressive spinal forms of multiple sclerosis and other chronic myelitis such as sarcoidosis, lupus, mixed connective tissue disorders, carcinomatous meningitis, AIDS myelopathy, and syphilis (tabes dorsalis) as well as subacute myelopathies such as those due to vitamin B12 deficiency. Acute inflammatory myelopathies can be confused with HTLV-1-associated myelopathy, especially when they have a subacute presentation that could mimic the rapidly evolving forms of HTLV-1-associated myelopathy. They include transverse myelitis, focal herpes virus myelitis, and paraneoplastic necrotizing myelitis. HTLV-1 can be associated with, and be the etiologic factor of, Sjögren disease, which can originate with a SICCA syndrome or a chronic myelitis. When both HTLV-1 and Sjögren disease are associated, it is not simple to untangle the real etiology (25). However, the result of treating with steroids and immunosuppressive therapy would be dramatically beneficial in Sjögren disease (55) compared to an absence of effect in HTLV-1-associated myelopathy.

When the patient belongs to an ethnic group known to be at risk, or comes from or lives in an area where HTLV-1 is endemic, the problem is different. In the West Indies, the carrier state affects 3% to 5% of the population, and in west coast Natives of the Pacific Northwest, it is evaluated at 2.5%. In some islands of Southern Japan, the carrier state was up to 20% of the population. In such circumstances, 20% of all unrelated neurologic problems might also occur in seropositive individuals. Because the frequency of HTLV-1-associated myelopathy among carriers increases with age, there is a high probability of coexistence of nonrelated pathologies such as compressive myelopathies, cervical arthrosis, stroke, and others.

This author’s experience based on Pacific Northwest native populations of British Columbia reveals that many native patients finally diagnosed as having HTLV-1-associated myelopathy had erroneously been diagnosed in the past with multiple sclerosis, progressive inflammatory myelopathy, cervical arthrosis, osteoporosis and compressive fractures, and "chronic rheumatic condition.” Conversely, some First Nation patients in Canada ended up not having HTLV-1-associated myelopathy. However, as they were carriers of HTLV-1 and also had an associated neurologic problem, they were wrongly diagnosed as HAM/TSP. These included diabetic neuropathy (1), syringomyelia (1), and stroke (2). In two patients, the author concluded that there probably was a coexistence of both HTLV-1-associated myelopathy and cervical arthrosis (48). Under these difficult circumstances, it becomes essential to demonstrate the existence of antibody synthesis intrathecally to affirm the diagnosis of HTLV-1-associated myelopathy (23). More specific, however, is the viral load in peripheral blood lymphocytes, which is higher (by at least 1 log10 or so) in patients with HTLV-1-associated myelopathy than in carriers (38).

Diagnostic workup

Simplified diagnostic criteria should be used only for epidemiological purposes and should not be applied clinically to identify HTLV-1-associated myelopathy and tropical spastic paresis. The epidemiological criteria include the following:

(1) presence of spasticity
(2) spasticity not present during childhood
(3) presence of bladder dysfunction; and
(4) disease of a progressive nature.

Clinically, these four elements may diagnose the disease as “possible.” Serological confirmation makes the diagnosis "definite.” These criteria are sufficient for epidemiological purposes, but applying them to individual patients for clinical purposes may result in errors in diagnosis with disastrous consequences.

In the presence of the appropriate clinical picture, in the individual patient and especially in nonendemic areas, the diagnosis should rest on an abnormal CSF, a relatively bland head MRI contrasting with thoracic cord atrophy, and antibody positivity with or without virus amplification by polymerase chain reaction.

Sohler and colleagues found that CSF pleocytosis in patients with HTLV-1-associated myelopathy is seen more often than previously reported (65% of their cohort) and is associated with shorter duration of disease and with the presence of MRI T2 hyperintense lesions in the cervical spine.

Many researchers are exploring different biological markers to differentiate between patients with HTLV-1-associated myelopathy and asymptomatic carriers. For example, profiling of HTLV-I Ag-specific Ab responses might prove a useful diagnostic tool, as was found in a cohort from Jamaica, where anti-Gag and anti-Env Abs were associated with disease status and were suggested to be used as a classification tool (16).

Blood smears can be useful when they reveal the presence of lymphocytes with multiple nuclei described as "flower cells.” These cells are found in HTLV-1-associated myelopathy, but also in carriers. Again, these cells, which are diagnostic of HTLV-1, cannot be recognized by automatic cell count where they are counted as polymorphonuclear leukocytes and, thus, must be identified on slides.

Slowing of somatosensory evoked potentials can confirm involvement of the spinal cord. In fact, slowing of vestibular evoked myogenic potential (VEMP) detects disease and confirms involvement of the spinal cord in the very early stages, before changes can be seen on MRI (17). Accordingly, it was suggested that patients who display altered VEMP responses should be closely monitored because they may develop neurologic diseases more frequently than do asymptomatic individuals with normal VEMP responses.

An MRI of the cervical and thoracic cord should be performed if only to rule out compressive myelopathies, vascular malformation, or syringomyelia. It could show the expected thoracic cord atrophy, which is seen in only 50% of patients (37), although it has been reported in up to 74%. Other less sensitive findings include atrophy and increased T2 signal in the thoracic cord and changes in the cervical cord (07). Swelling of the cord, although rare, has been reported in the acute phase (70). The demyelinating myelopathy of multiple sclerosis demonstrates increased signal on MRI of the cervical cord more frequently than HAM/TSP, but multifocal white matter abnormalities in brain MRIs are more in favor of multiple sclerosis than HAM/TSP (26).

Increased signal areas on T2-weighted cranial MRI are seen in as many as 58% of patients. Although common, these abnormalities do not permit differentiation of patients with HTLV-1-associated myelopathy from HTLV-1 carriers and from age-related or vascular changes (26; 43). MRI of the brain does have the advantage of clearly helping to differentiate this condition from multiple sclerosis even though the demyelinating areas are compatible with what can be seen in multiple sclerosis (periventricular, subcortical, and posterior fossa). The abnormalities rarely fit the criteria of Barkhof and colleagues for multiple sclerosis; their criteria need more than nine lesions of which some should be subcortical and some periventricular, and some can be in the posterior fossa or the spinal cord (08).

We have evaluated how the Barkhof criteria perform when applied to patients with HAM/TSP in an attempt to differentiate white matter lesions seen on head MRIs of patients with HAM/TSP from those seen in patients with multiple sclerosis. Barkhof criteria include the following:

I. one lesion larger than 6 mm
II. three lesions larger than 3 mm
III. one periventricular lesion larger than 3mm
IV. one infratentorial lesion larger than 3mm

All of the criteria were found at a higher frequency in the patients with multiple sclerosis than in the patients with HTLV-1-associated myelopathy/tropical spastic paraparesis (p < 0.05), including two commonly used criteria: the Paty and the Fazekas criteria.

Table 1. Criteria Differentiating Cranial MRI of 10 Patients With HAM/TSP and 11 Patients With Multiple Sclerosis

This author suggests that early in the course of the disease T2 lesions predominate in the thoracic spinal cord, affecting both grey and white matter: posterior columns, posterior horns, and lateral columns. These early T2 lesions are frequently more than two segments long and are accompanied by swelling of the cord. Later, atrophy prevails.

Severity of disease is correlated with a greater number of white matter changes in the brain and more thoracic cord atrophy (26). Nonspecific findings that may be attributable to subclinical pathology, other neurologic disease, vascular abnormalities, or age-related changes limit the use of brain MRI as a first-line diagnostic modality, except to rule out multiple sclerosis. Confirmatory evidence was presented by Morgan and colleagues in 2007 (43; 07). Diffusion tensor imaging is the most promising in identifying white matter lesions in HAM/TSP (39).

CSF shows normal opening pressure. Classically, at the beginning of the disease, CSF often shows lymphocytosis (5 to 25 cells/cu mm), with normal or discretely increased proteins, increased IgG with an oligoclonal pattern on electrophoresis or isoelectrofocusing, and increased IgG ratio and daily synthesis. In our experience, of 12 clinically definite cases, only six had increased cells in CSF, and only seven had oligoclonal bands.

CSF serology for HTLV-1 and polymerase chain reaction can be used to confirm the diagnosis. But generally, the diagnosis is made on serological testing for HTLV-1 antibodies. Due to the need for testing large numbers of samples, specialized labs perform a 2-step process. In a second confirmatory step, ELISA positive sera are studied by Western blot. Diagnosis should be confirmed by PCR.

Screening test. Since 1988, the FDA recommends the use of an ELISA as a screening assay against whole virus. ELISA does not distinguish between HTLV-1 and HTLV-2. When only ELISA assays are performed, without attempt to distinguish between the viruses, it has become customary to record results as “HTLV-1/2 positivity.”

Confirmation step. Western blot is the serological "gold standard" used to confirm ELISA results. Positive confirmation requires reaction with a core protein and at least one glycoprotein from the envelope. In most patients with HAM/TSP, antibody levels in serum are high, or at least much higher than reported for healthy carriers. Negative samples will contain no band or no anti-p21e antibodies. Results are "indeterminate" if only one of the core or envelope proteins is present and if p21e is absent. Indeterminate and p21e negative westerns rarely seroconvert, but repeat testing should be done. If on repeat testing no new bands appear, the sample should be considered as indeterminate. Western blotting can also differentiate between HTLV-1 and HTLV-2.

Gene amplification and polymerase chain reaction. Polymerase chain reaction is the best technique available to confirm the presence of proviral DNA and distinguish between HTLV-1 and HTLV-2. It can be performed on DNA derived from peripheral blood cells or cerebrospinal fluid lymphocytes. Primers specific for p24 gag, pol, env, and p12 of HTLV-1 are available. Quantitative PCR is not always available and has not been well standardized, but the number of proviral particles per lymphocyte is a good indicator of the infection status, generally 10 times higher in HAM/TSP than in carriers. It can be done on CSF (57; 38; 20).

Positive antibodies in the CSF can be used to confirm the diagnosis of HTLV-1-associated myelopathy. However, the diagnosis is not clearly eliminated by a negative test, as the findings of a negative serology in the CSF could arise from defective testing (the lab having made dilutions of the CSF) as well as from the absence of CNS disease. Calculations of antibody titers as a ratio of IgG in CSF and in serum permit definite assessment of an intrathecal secretion of specific antibodies, confirming the local CNS immune reaction to the virus and most probably the local presence of the virus itself (58). It would, thus, appear that only a polymerase chain reaction that directly evaluates the presence of viral sequences can confirm the diagnosis. It has even been suggested that in acute bouts of this chronic disease, proviral load in CSF cells can increase up to 10 times the load found in peripheral blood (24; 20).

Management

No treatment has been demonstrated by double-blind placebo-controlled trials to be able to slow down the course of HTLV-1-associated myelopathy. None of the treatments discussed below has been approved by the FDA. Indeed, running a double-blind placebo-controlled trial in this condition is extremely difficult, due to economic reasons including cost and the third world status of the populations that are endemically affected by the HTLV-1 virus.

One should, however, applaud the creation of the “International Retrovirology Association” and support the actions of the association, which now permits us to focus questions and trials on a group of dedicated and highly informed members. One of the goals of the association is to set up clinical trials. Its first achievement has been to publish guidelines for the management of HAM/TSP in 2018 (05). For more information, please visit the following website: http://htlv.net/.

Spasticity. Baclofen® 30 to 60 mg/day and tizanidine hydrochloride 2 to 12 mg/day are useful. Repeated high-dose treatments with methylprednisolone (1000 mg intravenous monthly) may be of help in reducing spasticity and pain. In open, nonblinded, noncontrolled, nonrandomized studies, intravenous methylprednisolone 1 g/day for three days every third month appears to help in slowing the course of the disease and reducing pain (10). Having treated approximately 30 patients with monthly pulses of 1000 mg of methyl-prednisolone, this author had a similar experience and would recommend this regimen when placebo-controlled, randomized, and blinded trials are not available. A controlled trial showed a benefit of corticosteroids, but results did not reach statistical significance for its primary endpoint (79).

Physical therapy. A physical approach is highly recommended (63). Reports from Iran that transcranial magnetic stimulation improves spasticity in these patients for up to 30 days (a duration similar to the observed effect of methyl prednisolone pulses) needs to be repeated in a properly controlled, randomized trial

Bladder care is of paramount importance and should start with the use of anticholinergic agents such as oxybutynin chloride 5 to 30 mg/day or tolterodine tartrate 2 to 8 mg/day. The use of anticholinergics seems counterintuitive, but bladder retention is caused by reduced proprioceptive input (similar to what occurs in tabes dorsalis), and anticholinergic medications will decrease the sphincter tone as well as the bladder tone. The Credé maneuver and forward abdominal flexion with the fist above the bladder can be useful. Clean self-catheterization is helpful. It is essential to protect renal function when post-void residual is greater than 500 ml.

Outcomes

Despite treatments that would appear to have a chance of being effective, most patients will have a protracted progressive worsening course, with progressive increase in spasticity and motor deficit resulting in death due to intercurrent UTI or chest infections. In the slow progressive form of the disease, the course follows decades of progression. In rapidly progressive courses, death occurs in a matter of a decade. Acute T-cell lymphoma, leukemia, is reportedly a cause of death in 3% of these patients (45).

Special considerations

Pregnancy

It is not known if pregnancy affects the course of the disease. The disease does not appear to affect pregnancy. The virus can be transmitted vertically. It is thought that there is no transplacental infection but, rather, that the infection is transmitted by breastfeeding. In populations at risk for the disease, it is advisable to screen pregnant women for HTLV-1 serology (50) in order to be in a position to discuss risks of breastfeeding. The Japanese Health Services have demonstrated that the risk of transmission of the virus disappears if infected women avoid breastfeeding, with immediate reduction in seropositivity in the next generation.

Breastfeeding. It is essential during pregnancy for HTLV-1-positive mothers to avoid breastfeeding. Indeed, this systematic approach has been very successful, leading to the near eradication of the disease in Japan in one generation.