Timothy C. Hain, MD. • Page last modified: March 4, 2021

Video of opsoclonus in young woman, developed after the West Nile outbreak in Chicago. See the site DVD page for a list of more movies like this one.

Opsoclonus denotes chaotic back-back saccadic eye movements. It is a dramatic syndrome, sometimes due to cancer or a brainstem encephalitis such as West-Nile or Dengue. There is an immense literature about opsoclonus, probably because it is so dramatic. Kinsbourne in 1952 described irregular twitching of the extremities and face accompanied by constant eye movements. David Cogan described it more precisely in 1954. This was followed by J. Lawton Smith and Walsh, in 1960. Cogan described the post-viral version of opsoclonus in 1968, and Davidson reported on the association of opsoclonus with neuroblastoma (in children) in 1968 as well.

Opsoclonus in young children is drastically different than in older persons -- we think it best to divide up opsoclonus by the age group - -pediatric, or adult at least. Often the pediatric neurologists are unaware that opsoclonus in adults and children differ, and make sweeping generalizations about cause and treatment, that are baseless. Armingue et al (2019) stated, quite reasonably, that "the most important indicator of paraneoplastic OMS is the patient's age". They were studying adults, and for adults under the age of 40, the probability of a tumor other than an ovarian teratoma, is very low.

Opsoclonus differs from ocular flutter in that opsoclonus changes rapidly in any direction (horizontal, vertical, torsion) -- i.e. the eye movement vector is chaotic, while flutter is generally always purely horizontal. Opsoclonus may look like a a chaotic "shimmer" on direct observation.

Ocular flutter is often missed by clinicians who lack the ability to view the eyes with a video system. Here we are somewhat liberal about lumping together opsoclonus with flutter, as we view them as being variants of the same underlying process (back-back saccades). In addition, it seems to be difficult for many clinicians to tell them apart.

Opsoclonus and flutter is generally best visualized by using a video-frenzel goggle system, having a large screen. Opsoclonus may be difficult to record -- and in other words, sending the patient to your local ENT for an VENG may not help you very much. A general rule is that the bandwidth (samples per second) of the device that you are using to record an eye movement should be at least twice the bandwidth of the eye movement. Because opsoclonus occurs so quickly -- that it cannot be captured very well by low-bandwidth devices such as clinical EOG or VNG systems. Furthermore, inexperienced individuals without any training in neurology often mistake opsoclonus or flutter for other more benign entities, such as square wave jerks.

In all age groups, opsoclonus is rare. The common causes of opsoclonus depend on age.

Pediatric opsoclonus:

The causes of opsoclonus in children are drastically different than in older groups. Opsoclonus in children is often caused by a neural tumor (a neuroblastoma). Rotherberg reviewed this subject in 2009. The median age in children is about 18 months. Hasegawa et al (2014) reported that in roughly half of their 23 patients, neuroblastoma was the etiology. Looking at it from the other side, according to Morales, about 2-3% of children with neuroblastoma develop opsoclonus (2012). About half of all referrals of opsoclonus is from parents watching online videos !

Opsoclonus is said to occur in 1/200 children with Neuroblastoma. The Neuroblastoma/Opsoclonus combination is actually a somewhat fortunate one for the patient, as while the long term survival is only about 50% in "all comers" Neuroblastoma, generally no mortality is reported in the group with opsoclonus(Boltshauser et al, 1979). In theory, opsoclonus in neuroblastoma is caused by infiltration of the tumor by B cells, or some other mechanism of immune stimulation. In this group of neuroblastoma patients, there is lack of amplification of N-Myc (which confers high mortality), there is lymphoid follicles in the tumor. Furthermore, removing the tumor does not halt the CNS response, and relapses reflect immune activation. These children worsen with intercurrent illnesses, suggesting that activation of their immune systems causes some bystander problems.

Curiously, the prevalence of neuroblastoma has increased progressively over time, being only 8% in the 1970's, and 43% in the 2000's. Presumably this relates to better imaging technology. (Brunklaus et al, 2011). According to Pang et al (2012) the prevalence of pediatric opsoclonus-myoclonus is very low -- only about 1/5 million.

Diagnosis:Regarding diagnosis, relatively little is known presumably because these patients are so rare.

CT/MR imaging has the highest detection rate (Brunklaus et al, 2012), but there is a gradual shift to using MRI instead, to avoid the high radiation of high resolution CT. Imaging is often falsely negative (in 50%). The majority are not detected by abdominal ultrasonography or chest X-ray. MIBG scan rarely detects neuroblastoma in children with opsoclonus, because these tumors are not very metabolically active. Similarly, VMA and HVA tests are usually negative. It is difficult to study these patients as they are so rare (Pike, 2013)

Opsoclonus is very dramatic and very easy to spot at the bedside -- once you see a single case, you never forget. It is not known if these patients also have the increased nystagmus under closed eyelids as is seen in the post-viral syndromes in adults (Zangemeister), but they probably do as one maneuver to elicit opsoclonus in children is to ask the child to close their lids halfway (or hold eyelids and blow air on face). Yonekawa suggested that these patients also have an enhanced startle response (2011).



Long term prognosis of pediatric opsoclonus with neuroblastoma:

In a review from 1990, no child had an IQ > 90. A more recent study showed children were "in the normal range', but with a DQ <100. Untreated children do much worse. (Mitchell et al, 2005)

Paraneoplastic syndromes -- opsoclonus associated with another tumor -- also occur in children (see comment below about adults). (Singhi et al, 2014). Kruer et al ( 2014) reported a pediatric case with GABAB antibodies. This case was dominated by intractable seizures.

Occasional reports exist of rotavirus induced opsoclonus (Gurkas et al, 2014)

Rare cases have been reported from "vanishing white matter disease" -- a genetic syndrome (Klingelhoefer et al, 2014), followed by schizoaffective illness.

Adolescent opsoclonus

In adolescents through roughly the 60's, generally no cause is found and opsoclonus is blamed on the usual mysterious suspects -- viruses, autoimmune disorders and genetic defects. In other words, neuroblastoma is not a significant cause of opsoclonus once you get to be about 12.

Of course when this syndrome follows a viral infection such as a cold, it is difficult to be sure that this is more than a coincidence. Adolescents have more psychiatric disorders than other age groups, and it is possible that in some cases these are teens who have learned to produce an unusual variety of voluntary nystagmus. Adolescent opsoclonus often resolves without any treatment, after the child is kept out of school and tutored or home schooled for a year.

Kang and Kim reported a case in mumps (2014). Occasionally immunization for HPV precedes opsoclonus (McCarthy and Filiano, 2009).

These patients are not treated as are the pediatric cases (i.e. 2 year olds), with heavy immune suppression, but instead more conservative management prevails. In our clinical experience, this is generally with good results after a year.

Opsoclonus in older persons

Opsoclonus in adults is very different than opsoclonus in children, and they should not be lumped together. In children (largely 2 year olds), neuroblastoma is the main cause. In adults, generally there is no cause established, and neuroblastoma is almost never implicated.

Moving on to the data, Klaas et al (2012) reviewed 21 Mayo clinic patients as well as 116 previously reported patients with opsoclonus-myoclonus in adults. The median age was 47. More than half of the patients reported dizziness and imbalance. There were many causes, most of which were essentially mysterious (called "parainfectious").

Our clinical experience is that adults with opsoclonus or similar illnesses (e.g. flutter with ataxia) rarely have myoclonus as is reported in the childhood version of this illness. Thus the mnemonic "dancing eyes -- dancing feet", while easy to remember, is rarely appropriate. There are very rare reports of enhanced startle (Sotrichos et al, 2011). We have not seen this.

Smith et al (2010) reported neurocognitive decline in a single 52 year old patient.

"Parainfectious opsoclonus", possibly also including autoimmune as well as mistaken inference.

Infectious causes of opsoclonus/flutter
Organism # reports
West Nile 5
Dengue 3
HIV-1 5
enterovirus 1
hepatitis-C, hepatitis-A 2
HSV-6 1
Influenza-A 1
Lyme 4
Typhus 2
Varicella-Zoster 1
Mycoplasma 2
Streptococcus 1
Lyme 2

Before beginning the discussion, it is possible that most of these reports are just wrong. For example, Turner et al (2018) reported on "a case report of sphenoid sinusitis causing opsoclonus myoclonus syndrome". How do they know ? They are reports of associations, not of proven cause. Perhaps someone, for example, just happened to have the flu, and also developed opsclonus from some other reason. There is no way to tell.

Similarly, the table above may be meaningless. There are huge differences in the degree to which diseases are reportable, and also the risk of developing a disease.

Well anyway, starting with the more common "parainfectious" group, opsoclonus may also be caused by a viral infection of the brainstem or cerebellum, as well as autoimmune processes. A huge surge in opsoclonus/flutter cases appeared in Chicago, after the West Nile Virus outbreak of 2003. Reports have continued in the literature (e.g. Cooper and Said, 2014 see WNV page for more) It has now vanished again, as have most of the West Nile cases. There have also been several reports of opsoclonus after dengue virus infection (e.g. Wiwanitkit, 2014). Note that West Nile and Dengue are both members of the flavivirus family.

There are sporadic reports in other viruses as well such as enterovirus-71 (Akiyama et al, 2008), hepatitis-C (Ertekin et al, 2010), hepatitis A (Lee et al, 2019), HSV-6 (Belcastro et al, 2014), HIV-1 (Vale et al, 2013; Wiersinga et al, 2012; Klaas et al, 2012; Kanjanasut et al 2010; Scott et al, 2009), influenza A (Morita et al, 2012), scrub typhus (D'sa et al, 2012), Varicella-Zoster (Singh et al, 2010). In HIV-1, the reports mainly report to initial infection.

A few non-viral cases have been reported associated with Mycoplasma (Nunes, 2011; Huber et al, 2010), Streptococcus (Dassan et al, 2007), and Lyme (Skie et al, 2007;Peter et al, 2006). Scrub Typhus and Typhus, are from rikettsial diseases, which are not viruses but a variant of bacteria.

So essentially, after WNV and HIV-1 which have many reports, there are a lot of "one-off" viruses reported and occasional bacterial reports. Given the lack of effective treatment for most of these organisms, it is difficult to be very enthused about doing viral testing except perhaps in HIV-1, but it does have the value of making the neoplastic variant less likely.

Patients with this variety of opsoclonus often develop wild nystagmus under closed eyelids, which can be appreciated by watching them with eyes closed (Zangemeister et al, 1979). This is an excellent neurological sign. Nevertheless, opsoclonus is usually a "wastebasket" type diagnosis, arrived at after screening for cancer has been unproductive.

Occasional patients have autoimmune thyroid disease (Kuwahara et al, 2013; Salazar et al, 2012), and rarely patients have antibodies to GAD (Bhandari et al, 2012; Marakis et al, 2008), or NMDA (Kurian et al, 2010). Celiac disease is another rare association (Wong, 2007)

Treatment is rarely possible as most of the viral illnesses lack treatment. Some clinicians report using IVIG (Nunes et al, 2011), but it is difficult to know if this treatment was better than placebo.

Paraneoplastic opsoclonus

Armangue et al (2016) reported on 114 adult patients. Of these, 39% had paraneoplastic opsoclonus, and the remainder had what they called idiopathic OMS (opsoclonus). According to Klaas et al at Mayo (2012), only 3 of their 21 adult patients had cancer, presumably meaning that the rest were either undiagnosed or attributed to viruses.

Opsoclonus due to cancer is almost never due to tumor in the brain, but instead is generally attributed to a paraneoplastic syndrome (a tumor elsewhere in the body). In older adults, a full-scale workup for neoplasm is generally indicated and is productive in persons with opsoclonus. The frequency with which tumors are found in the less intense version of opsoclonus, ocular flutter, is so far unreported, but our estimate from our frequent appreciation of this in our clinical practice is that cancer as a cause is extremely uncommon.

Lung cancer, especially small cell, is the commonly found tumor (Laroumange et al, 2014). Thus a chest-Xray or CT scan of the chest (rather than an MRI of the brain), and a mammogram in women, are usually the most productive first tests. We ourselves in our clinical practice have encountered opsoclonus primarily in lung or breast cancer patients. There are sporadic reports in many cancers, including breast cancer (Weizman and Leong, 2004), squamous cell cancer of the esophagus (Rosser et al, 2014), pancreatic cancer (Nwafor et al, 2019), squamous cell cancer of the thymus (Yamaguchi et al, 2013), seminoma of the testicle (Newey et al, 2013), gastric cancer (Biotti et al, 2012), endometrial cancer (2010), Non-Hodgkins lymphoma. Wong (2007) also noted that there were cases of renal adenocarcinoma.

Thus in essence, after checking for the easy ones (lung/breast), one needs a wide screen to exclude occult cancers. PET scanning has also been used to diagnose occult cancers in this situation (Bataller et al, 2003).

Although antibodies such as anti-Hu, Yo, and Ri among others may occasionally be positive (generally considered as a screen for neoplasms), commercial testing for antibodies is often of little diagnostic value. Armangue et al (2016) found "Onconeuronal antibodies occurred in 13 [of their 114] patients (11%), mostly Ri/ANNA2 antibodies, which were detected in 7 of 10 patients (70%) with breast cancer. "

Research studies have implicated autoantibodies to a large assortment of miscellaneous neural antigens (Blaes, Fuhlhuber et al. 2005; Panzer et al, 2015; Player et al, 2015). Sabater et al (2008) suggested that antibodies are heterogeneous in adults, unlike the situation with children with neuroblastoma. As an example, occasional patients have CSF antibodies to GABAB (DeFelipe-Mimbrera et al, 2014), and GQ1b (Zaro-Weber et al, 2008). To summarize, we are dubious as to the clinical utility of antibody testing in adults with opsoclonus.

It is important to watch for signs of worsening in patients with saccadic disorders, as of course, tumors related syndromes rarely regress. This is not 100% true, there are some reports of regression in tumor patients (Simister, 2011). We ourselves have seen regression in patients who were treated for their cancer.

Voluntary nystagmus mistaken for opsoclonus

Adults, like adolescents, sometimes learn to produce voluntary nystagmus, usually to get attention or perhaps to obtain disability payments. Generally speaking, the pupil constricts when they are doing this, and of course, there are no neurological findings.

Other causes.

Diphenhydramine poisoning is sometimes associated with opsoclonus. (Irioka et al, 2009; Herman et al, 2005). As diphenhydramine is available "over the counter", this is something to consider in individuals who might be abusing medication.

Mechanism of opsoclonus.

Opsoclonus is classically attributed to malfunction of the pause-cells in the midline brainstem. However, this is probably wrong. The first hint of this was a report of Ridley (1987) who examined pause neurons in autopsy studies and noted that they were normal in two cases of opsoclonus associated with small cell cancer. Of course, 2 cases is hardly a comprehensive study of opsoclonus.

Further studies have shown abnormalities mainly in the deep cerebellar nuclei. PET scan studies show activation in the deep cerebellar nuclei in opsoclonus and ordinary MRI sometimes shows lesions (Helmchen et al, 2003; Boland et al, 2012; Mustafa et al, 2015). Similar activation in the deep cerebellar nuclei was reported in a case of Newey et al (2013). The fastigiate oculomotor region projects to the burst neurons, omnipause neurons, and the local feedback loop of the brainstem saccade generator. The fastigiate oculomotor region is inhibited by the vermis. Thus the fundamental underlying problem in opsoclonus may relate to decrease activation in the posterior vermal lobule VII.

Others have reported changes in the thalamus, hypothalamus, and pons on Flair (Chen et al, 2012). Tsutsumi et al (2009) reported a case with a lesion in the right upper pontine tegmentum, including the superior cerebellar peduncle, and suggested that the lesion interrupted fastigial connections.

Ramat et al (2008) reported that they could make their mathematical model oscillate by reducing the inhibitory effect of pause neurons. Shaikh et al (2008) suggested that "ion channel dysfunction in the burst cell membrane" in the underlying abnormality. While perhaps true, the question then is what is the source of the ion channel or pause neuron dysfunction ?

Treatment of opsoclonus

Treatment of opsoclonus depends on the cause. Basically we have paraneoplastic, autoimmune, infection related (? autoimmune as well), and idiopathic.

When it is paraneoplastic, treatment is focussed on removing the tumor. Attempts made to treat with immunosuppressants are often unsuccessful (Hassan et al, 2008), but some responses are reported (Ohara et al, 2007)

In very young children with opsoclonus there is a vigorous effort to find and remove a potential tumor as well as treatment with powerful immune suppressant medications. (Toyoshima et al, 2015). As mentioned above however, in children, the neuroblastoma is generally not very aggressive, and it doesn't matter much if the tumor is taken out or not, so to some extent finding and removing the tumor is low yield.

Regarding treatment of the autoimmune variety, expect that immunosuppression will be needed for 5-7 years. There is some difference as to how this is done.

Pranzatelli et al (2013) proposed treatment with powerful immune suppressants such as Rituximab, steroids, or chemotherapy for opsoclonus including those with neuroblastomas. This is almost never done in adults with similar symptoms.

Hero et al (2013) states that the treatment is largely steroids.

Rituximab reduces activated B cells for 6-9 months and reduces use of sterodis. Generally speaking, children need 2-3 courses of Rituximab at intervals of about 6 months

Tate et al (2012) suggests that corticotropin should be combined with other immunosuppressants. Pranzatelli has also suggested ACTH may be used (2012). So it does appear that ACTH is reasonable as of 2017. After treatment for opsoclonus associated with neuroblastoma, most have persistent neurological sequelae (Krug et al, 2010; De Grandis et al, 2009)

Wilbur et al (2019) examined a "brief upfront .. therapy with rituximab" in a total of 7 children treated, and 8 control patients. This was an add-on. They reported that rituximab allowed reduction in the duration of steroid and IVIG.


When opsoclonus is post-viral or the tumor is gone, or just idiopathic, then treatment is symptomatic. Strupp et al (2006) comment that disorders such as ocular flutter and opsoclonus are "still difficult to treat". (Strupp and Brandt, 2006)

Symptomatic treatment of opsoclonus largely consist of medications that slow down rapid neuronal firing. Examples include gabapentin, clonazepam(Paliwal et al, 2010; Bartos, 2006), and oddly enough, diphenhydramine (which is also reported to cause opsoclonus). Supposedly benzodiazepines or diphenhydramine can also worsen symptoms. We find this a bit dubious.


Case reports organized by infectious agent:

Chikungunya and dengue together



Hepatitis A or B



HSV and Zoster









Please read our disclaimerReturn to Index. • Page last modified: March 4, 2021