Clinical manifestations

Presentation and course

Definition. The First International Scientific Meeting on Nodding Syndrome held in Kampala, Uganda, in 2012 approved the first standardized case definition for nodding syndrome (67). Nevertheless, this case definition is intended to be reviewed and modified as new data become available (20). Additionally, consensus was reached to denominate the combination of signs and symptoms as nodding syndrome and replace the terms “nodding disease” and “head nodding disease.”

Suspected case.

• Head nodding is reported in a previously normal person. • Head nodding is defined as repetitive, involuntary drops of the head to the chest on two or more occasions.

Probable case. Suspected case plus both of the following major criteria:

• Age at onset of nodding between 3 and 18 years old • Frequency of nodding 5 to 20 per minute

And at least one of the following minor criteria:

• Other neurologic abnormalities (cognitive decline, school dropout due to cognitive/behavioral problems, other seizures or neurologic abnormalities) • Clustering in space or time with similar cases • Triggering by food and/or cold weather • Stunting or wasting • Delayed sexual or physical development • Psychiatric symptoms

Confirmed case. Probable case plus a documented nodding episode observed by a trained healthcare worker, or videotaped, or on EEG/EMG.

In 2008, Winkler and colleagues proposed and established a classification system that separates patients into subgroups with head nodding episodes only (HN only) or with further forms of epileptic seizures (HN plus) (63). A Ugandan study on the neuropsychiatric aspects of nodding syndrome proposed to classify the syndrome into “neurologic nodding syndrome,” “psychiatric nodding syndrome,” and “mixed nodding syndrome” (38).

Nevertheless, reports about the clinical features and course of nodding syndrome may differ, and it remains uncertain whether the reports about nodding syndrome all refer to the same disorder. In the long-term course, physical and cognitive decline may differ in the different regions. Compared to reports from Uganda and Sudan, children with nodding syndrome in Tanzania seem to be less severely affected, and, hence, prognosis may be better (63; 64; 66). Current evidence and surveillance data suggest that death mainly results from secondary causes, like drowning or burns from fire, due to seizures or malnutrition (Ugandan Ministry of Health, unpublished).

EEG and semiology. Nodding syndrome is most likely an epileptic encephalopathy (58; 66; 50; 07). Tumwine and colleagues described a series of repetitive, involuntary dropping of the head, lasting 2 to 5 minutes with a frequency of 10 to 20 per minute (58). The episodes were noted to happen more frequently in the morning or while eating. Although some children were unresponsive to commands along with absently staring, consciousness was not impaired in others as they continued to eat and follow commands to rise, walk, or turn. Sometimes such nodding episodes evolved into generalized tonic or tonic-clonic seizures; in others, episodes of sudden shouting, screaming, jumping up, and running in circles were reported by caretakers. Tumwine and colleagues described another patient in whom head nodding with preserved responsiveness could reliably be induced by local food prepared from maize (ugali), although it was not observed when the patient ate a Western candy bar (58). Similar semiologic descriptions have been also reported by others (63; 18; 50).

To date, several studies have reported ictal and interictal findings on EEG. Idro and colleagues reported abnormal routine EEGs in 20 of 22 children, with abnormal interictal findings, which included slow background activity as well as generalized and focal epileptiform discharges, suggestive of presence of symptomatic generalized and focal epilepsy (18). No prolonged or overnight EEGs recordings were performed and ictal findings were not reported. Routine video-EEGs in a study by Sejvar and colleagues showed abnormal interictal activity in 10 of 12 children with frequent 2.5 to 3 Hz generalized spike-and-wave discharges, multifocal spikes, and polyspikes (50). Complex partial seizures were recorded in two children; in one of the children, seizure onset was preceded by consumption of a cold beverage. Nodding episodes were recorded in two children, which electrographically demonstrated a generalized electrodecrement followed by generalized sharply contoured rhythmic theta activity, which were categorized as atonic seizures. Soldatos and colleagues performed long-term EEG monitoring in three individuals with nodding syndrome from the same family (53). They reported interictal patterns of sleep-activated bursts of generalized and multifocal epileptiform discharges in two of three patients. The third individual had diffuse background slowing and multifocal epileptiform discharges without sleep activation. For this individual, 16 clinically similar ictal events were noted with altered awareness and bilateral limb tonic-clonic movement. Electrographically, the onset varied with diffuse attenuation that evolved in the temporal region (either left, right or bitemporal). Additional seizures with different semiology—asymmetric bilateral upper extremities elevation with associated clonic movements—are followed by bilateral tonic-clonic seizures and brief jerks of all four extremities. Ictal EEG for these seizures showed right posterior quadrant onset and bilateral posterior temporal onset (53). Tumwine and colleagues described routine EEGs of 32 subjects with nodding syndrome showing universal recurrent paroxysmal pathological discharges (58). In three cases, nodding episodes were recorded showing paroxysmal slow-wave rhythms “followed by a brief and small fast discharge.” This activity was observed in a pseudo-periodical manner for several minutes.

In contrast, Winkler and colleagues reported unremarkable routine EEGs in four (40%) of 10 patients with head nodding (one with HN plus); abnormal background activity and intermittent sharp slow-wave activity were reported in two patients (63). Seizures or nodding episodes were not recorded. In their follow-up study, Winkler and colleagues reported abnormal background activity in 11 (44%) of 25 patients (66). Generalized ictal 2.5 Hz spike-wave activity was recorded in two patients (8%), and interictal bursts of 1.5 to 2 Hz spike-wave rhythms were observed in 4 of 25 (16%) patients.

de Polo and colleagues reported abnormal EEG findings in 20 of 21 children with nodding syndrome from South Sudan (07). Interictally, background activity in majority of the children was found in the delta and theta range. Eighty-five percent of the children showed bilateral 2 to 3.5 Hz spike and wave epileptiform discharges over the fronto-temporal or fronto-centro-temporal regions (75%) and fronto-parietal-temporal region (10%) that were often intermingled with sharp waves. Ictal EEG in three children with nodding syndrome showed a high-amplitude, bi- or triphasic slow wave with inverse phases reversal over the vertex region. These slow waves were occasionally followed by an electrodecrement with diffuse fast activity for 2 to 4 seconds. The ictal activity repeated every 3 to 10 seconds and often occurred in clusters, correlating clinically, with nodding episodes (07).

Mazumder and colleagues reported the ictal electrographic features of nodding syndrome patients from the three highly prevalent regions of Uganda, South Sudan, and Tanzania (35). Confirmed nodding syndrome cases from the noncontiguous regions underwent video-EEG, which showed slow waves followed by electrodecrement and superimposed gamma activity during ictal head nodding episodes. The combination of ictal EEG with electrodecrement and clustering of head nodding episodes were similar to electrographic findings of late-onset epileptic spasms, suggesting that nodding syndrome is a unique form of childhood onset epilepsy.

Although there is a general consensus that head nodding events are atonic seizures, given the heterogeneity of the neurophysiological data, EEG is not yet considered an essential criterion for diagnosis of nodding syndrome (67).

Clinical course. In a large retrospective study of 210 children with nodding syndrome in Uganda, Idro and colleagues identified five overlapping clinical stages with deteriorating epilepsy, cognitive impairment, and functional and psychiatric abnormalities (19).

Prodromal stage. An initial prodrome with “dizziness” and loss of attention, sleepiness, lethargy, and staring was reported in four patients up to 6 weeks prior to the first head nodding episodes.

Head nodding. This stage was found to be the pathognomonic feature of the syndrome. The mean age of head nodding was 7.5 years. Characteristic dorso-ventral head drops were noted with frequency of 5 to 20 Hz lasting several minutes. With progression of disease there is development of other types of seizures including atypical absences and myoclonic jerks.

Convulsive seizures. Within 6 months to 3 years, children with nodding syndrome develop convulsive seizures, cognitive dysfunction, and psychosis.

Functional impairments. Around 2 to 5 years after head nodding spells appear there is significant functional impairment. During this time there is gradual deterioration in behavior, motor skills, speech, cognition, and psychiatric symptoms. Some patients developed skeletal abnormalities such as kyphosis and limb and pectus deformities. The children who were still active and mobile, wandered around or ran away, and were prone to getting lost.

Severe disability. Children lose their mobility and appear severely wasted and apathic, with flat affect, limited speech, and poor appetite; some develop limb deformities.

Reliable long-term data are not available, and whether wasting (stage V) is obligatory in the course is still controversial and is only infrequently seen in Tanzania (66). Sejvar and colleagues followed 12 of 23 children for 8 months without seeing improvements in nodding or seizure frequency (50). Over a period of 4 years, Winkler and colleagues followed 53 of 62 patients, the majority receiving treatment with antiepileptic drugs (66). Among the 23 patients initially classified as head nodding only, 10 remained head nodding only and five deteriorated towards head nodding plus with additional generalized tonic-clonic seizures. Head nodding ceased in two patients with later development of other seizure subtypes. In six of the original head nodding only patients, head nodding stopped; however, they developed other seizure subtype. Among the 30 initial head nodding plus patients, nine remained head nodding plus with head nodding and other seizures. In 15 patients, head nodding ceased although the other seizures persisted; four patients with head nodding plus changed to head nodding only, and head nodding and generalized tonic-clonic seizures in two patients had stopped by follow-up.

Overall, head nodding stopped in 25 of 53 (47.2%) interviewed patients, with a median head nodding duration of 5 years at a median age of 15 years, although 21 patients (84%) developed other epileptic seizures, mainly generalized tonic-clonic seizures, and 20 of the 25 (80%) in whom head nodding had stopped were still on antiepileptic drugs. Four of 53 (7.6%) patients (two previously classified as HN plus) were seizure free, with head nodding seizures stopping at a median age of 13 years; two were still on antiepileptic drug treatment.

Neurologic deficits and functional disabilities. Olum and colleagues assessed the neurologic and functional disabilities of nodding syndrome patients 10 years after onset (43). The median age of onset was 8.4 years with a mean age of 18.9 years in the affected children. Twenty eight of 57 patients were severely functionally disabled and likely to have neurologic and cognitive impairment. The affected children were characterized as mute or near mute, exhibiting drooling and having limited expression or animation. About one fifth of children suffered severe burns due to falls or seizures and were frequently stated to be unable to feed themselves or perform household chores. However, 15 of 57 patients were described as well functioning and were able to perform daily activities such as fetching water, digging, washing dishes, or playing football. Affected children were described as “slow” and were likely to repeat school for several years or drop out at an early age. Focal neurologic abnormalities were identified in 28 of 57 patients, with the most common abnormality being increased tone. Twenty eight of 57 patients developed pyramidal signs such as hyperreflexia, positive Hoffman sign, clonus, or extensor plantar responses. Increased tone was observed in one limb (n = 7), two limbs (n = 8), three limbs (n = 3), and four limbs (n = 11). Spasticity in all four limbs was reported in all 28 children. Few patients (16%) were reported to have extrapyramidal signs. Despite pyramidal and extrapyramidal abnormalities, muscle strength was maintained in most patients. The absence of peripheral nervous system and cerebellar involvement suggests that nodding syndrome presents as a multisystem central nervous system disorder with pyramidal and extrapyramidal involvement.

Cognitive impairment. Severe cognitive impairment has been consistently noted, although standardized neuropsychological assessment has been infrequently performed in the rural setting in Africa. Severe cognitive impairment was reported in 15% and 40% of nodding syndrome patients in Sudan and Tanzania, respectively, although the standardized neurocognitive tests were not performed, and the patients were graded according to the parents’ assessments (64; 58). In Uganda, Sejvar and colleagues performed a brief self-designed, 5-item test and Mini Mental Status Examination (MMSE) to assess 65 children with nodding syndrome (50). Compared to age- and village-matched controls, the patients displayed a reduced ability to copy simple and complex shapes, reduced category fluency, reduced digit span, and impaired finger tapping of the dominant hand. Anecdotally, the cognitive impairment appeared to be progressive, although there was no long-term observation performed (58; 50). Idro and colleagues described cognitive impairment and school dropout within 2 to 4 years after the onset of head nodding in 22 children (18). Cognition was assessed in three children between 13 and 15 years of age using the KABC-II (Kaufman Assessment Battery for Children) test two weeks after beginning treatment with sodium valproate. The study demonstrated severe cognitive impairment in multiple cognitive domains. Working memory, planning, learning, visuospatial, and knowledge abilities were similar to children at age five or younger.

Psychiatric abnormalities. Children with nodding syndrome often show psychiatric abnormalities. Idro and colleagues reported four children with paroxysmal episodes of fear, panic, and visual hallucinations, with one child shouting and running at the onset of hallucinations and a fourth seeing a person with knives trying “to kill her” (18). Tumwine and colleagues also cited caretakers reporting similar episodes of sudden shouting or screaming, jumping up and running in circles, as well as agitation, weakness, daytime sleepiness, and fluctuating mental abilities (58).

Prognosis and complications

Previously, nodding syndrome was regarded as invariably fatal by local caretakers, although available data do not support such reports. During their 2012 census study, the Ugandan Ministry of Health reported 19 deaths among 767 patients with probable nodding syndrome (2.5%), although they provided no time reference or details (20).

At present, only one long-term study is available; Winkler and colleagues followed up with 53 of their original 62 patients after 4 years (65). Two patients (3.5%) had died, one during a prolonged status epilepticus and one from traumatic brain injury, with seven patients lost to follow-up. Additionally, there is significant socioeconomic cost for affected households (30).

Biological basis

Etiology and pathogenesis

The etiology of nodding syndrome remains uncertain, although the disorder emerged in areas of extreme poverty, recent food shortages, armed conflicts, and population displacement (58; 09; 54). Several observational case series and case-control studies demonstrated wasting and stunting in patients with nodding syndrome, but failed to consistently identify associations with consumption of several distinct foods, micronutrient malnutrition (cobalamin, folate, retinol, or selenium), or ingestion of toxic substances (mercury, homocysteine, thiocyanates, copper, lead, arsenic, pesticides, pesticides, or mycotoxins) (58; 09; 13; 18; 50; 57; 10). Spencer and colleagues proposed that malnutrition, B6 hypovitaminosis, and consumption of Agaricus bingensis (a hydrazinic mushroom) to be potential causes of nodding syndrome that call for greater investigation (55). It is not uncommon for families to rely on the consumption of A bingensis during periods of near famine. Hydrazine-containing compounds induce DNA damage that manifest as neuropathological changes like those found in nodding syndrome. Hydrazine has also been related to B6 hypovitaminosis and reduced levels of vitamin B6 have been associated with higher plasma levels of 3-hydroxykynureunine (3-HK), a neurotoxic compound linked to epilepsy. In Sudan, Tumwine and colleagues reported consumption of red sorghum, consumption of baboon brain, and absence of a previous measles infection as being associated with nodding syndrome, although Foltz and colleagues failed to reproduce these observations (58; 13).

In Uganda, Spencer and colleagues reported an epidemiological association of nodding syndrome with prior history of measles infection, which was confirmed in subsequent serological studies (54; 61). However, the authors used the presence of IgG as a measure of prior measles exposure, and thus, it remains unknown if this higher prevalence of measles antibodies was an indication of prior infection or higher rates of vaccination among the cases (61). On the contrary, Soldatos and colleagues did not depict any associations to prior measles infection or detectable measles RNA in CSF for three nodding syndrome patients in Uganda (53). Out of the three cases, only one case was determined to be positive for the hepatitis B and G virus in serum.

In South Sudan, Edridge and colleagues identified a divergent rhabdovirus, which they named Mundri virus (MUNV), in the serum but not CSF, of a 15-year-old girl with new-onset nodding syndrome (11). A case-control study of 72 patients with nodding syndrome was then initiated to determine the seroprevalence of MUNV, but no other patients were found to have detectable MUNV RNA in serum. Similar to previous findings, no association was found between nodding syndrome and rubella, HSV-1, or CMV viruses.

Considering inflammatory markers, Ogwang and colleagues conducted a case-control study with 150 subjects in each group and found elevated C-reactive protein (CRP) and C5/C5A in the sera and CSF, respectively, of subjects with nodding syndrome as compared to controls (42). The authors concluded this finding to be suggestive of a nonspecific inflammatory state of still undetermined origin that is possibly infectious, postinfectious, or autoimmune mediated.

Two case-control studies from South Sudan have found an association between nodding syndrome and parasitic infection with Mansonella perstans (56; 12). Of importance, Edridge and colleagues included cases of nodding syndrome with recent onset of symptoms less than one year ago and found the following infections: Mansonella p (odds ratio 7.04, 95% confidence interval 2.28-21.7), Necator americanus (OR 2.33, 95% confidence interval 1.02-5.3), and antimalarial seroreactivity (OR 1.75 95% confidence interval 1.2-2.57) (12).

Other infectious agents, such as malaria, trypanosomiasis, cysticercosis, prion disease, and hepatitis E, have not been found to be associated with nodding syndrome (09; 13). Although most studies have found persistent epidemiological associations with the microscopic helminthic parasites Onchocerca volvulus and Mansonella spp, other filariae such as Loa loa and lymphatic filariasis do not seem to be related to nodding syndrome (58; 09; 13; 57).

Onchocerciasis and nodding syndrome. Epidemiological associations between onchocerciasis and epilepsy in general are reported in several studies (45; 24). Similarly, compared to children without nodding syndrome, a higher prevalence of onchocerciasis was consistently demonstrated in children with nodding syndrome by skin snip microscopy and OV16- and OvFAR/OvMSA-based serology (58; 09; 13; 57). Studies from Tanzania did not find evidence for either the presence of O volvulus in CSF by polymerase chain reaction, nor evidence of O volvulus infection in the CNS by detection of IgM or IgG antibodies against the parasite in CSF, despite O volvulus antibody positivity in serum in 81% of the tested patients with epilepsy (27). This led to the possibility that nodding syndrome is an autoimmune-mediated disease caused by molecular mimicry with O volvulus antigens (22). However, in South Sudan, Edridge and colleagues studied cases of nodding syndrome with recent onset of symptoms less than one year ago and did not find any association between Onchocerca volvulus exposure and nodding syndrome (12).

Autoimmunity and nodding syndrome. Johnson and colleagues used unbiased protein chip methodology and detected autoantibodies to leiomodin-1, a neuronal cytoplasmic protein that is cross-reactive with O volvulus antigens in the sera and CSF of cases with nodding syndrome from Uganda (22). The authors found leimodin-1 antibody more commonly in the sera of nodding syndrome cases than in sera of the unaffected controls from the same community. Furthermore, the authors were able show that the antibodies to leiomodin-1 were neurotoxic in vitro. However, the same study reported the presence of leimodin-1 antibodies among the unaffected individuals from the same community who were O volvulus positive, and it remains unclear why these individuals do not show symptoms of nodding syndrome. A more perplexing issue was that only 50% of cases with nodding syndrome showed antibodies to leimodin-1 in the CSF. Although this study provided initial evidence to establish nodding syndrome as an immune-mediated epileptic encephalopathy, a causal relationship remains to be established. In addition, Soldatos and colleagues found evidence of prior O volvulus infection in three nodding syndrome patients, who were extensively evaluated at the National Institute of Health from the same family (53).

Conversely, Hotterbeekx and colleagues analyzed the sera of individuals with onchocerciasis-associated epilepsy including nodding syndrome versus healthy controls via a cell mediated and Western blot assay, and no association was found between Onchocerca volvulus infection or epilepsy status and the presence of leiomodin-1 autoantibodies (16). However, only 27% of the affected subjects had nodding seizures, the rest having onchocerciasis-associated epilepsy without nodding seizures. This contrasts with Johnson and colleagues whose study population only consisted of children with nodding syndrome (22).

Levite and colleagues studied 30 individuals with nodding syndrome and healthy controls from South Sudan and found that a significantly higher proportion subjects with nodding syndrome have antibodies to three extracellular peptides of glutamate receptors: anti-AMPA-GluR3B (86%), anti-NMDA-NR1 (77%), and anti-NMDA-NR2 (87%) (32). The same study also found that purified IgGs from subjects with nodding syndrome, particularly subject’s anti-GluR3B antibodies, killed both human neural cells and T cells. Furthermore, the authors were able to induce seizures in vivo after infusing IgG from subjects with nodding syndrome into the brains of normal mice. These experimental findings led the authors to conclude that nodding syndrome is modulated by autoantibodies to the GluR3B receptors that expressed on both neural and T cells, and these glutamate antibodies bind and injure both cell types. A dual-hit on the immune and the nervous system subsequently results in nodding syndrome and overall deterioration of general health due to an external environmental trigger such as Onchocerca volvulus.

Another study investigated the role of antineuronal antibodies against N-methyl-D-aspartate and voltage gated potassium channel receptors in pathogenesis of nodding syndrome in a Tanzanian cohort but did not find any association (08). The difference in the findings among the three3 cohorts from Uganda, Tanzania, and South Sudan needs to be further evaluated.

Genetic aspects. The high familial occurrence and clustering within circumscribed villages and tribes may suggest a hereditary nature of nodding syndrome, as well as of the associated generalized epilepsy and its specific genetic traits. Regarding the area around Mahenge in Tanzania where nodding syndrome seems endemic rather than epidemic, extraordinarily high epilepsy rates in adults and children have been reported--up to 37 per 1000 inhabitants (02; 49; 39; 34). Clusters of people with epilepsy were found in distinct tribes and villages, and within families, whereas other tribes and villages were hardly or not affected (58; 18; 50; 57). Patients with nodding syndrome often have first-degree relatives with epilepsy, and siblings are frequently affected (63). There is a paucity of detailed workup for genetic disorders, although exome sequencing in a child from Uganda and another from South Sudan has not revealed any disease-related changes (09).

Benedek and colleagues evaluated the role of human leukocyte antigen (HLA) class I and II molecules in the pathogenesis of nodding syndrome (04). In 48 subjects with nodding syndrome and 51 health controls from South Sudan, the authors examined seven human leukocyte antigen loci by next-generation sequencing or sequence-specific oligonucleotide probes. The study found both protective and susceptible haplotypes. Particularly, HLA-B*15:31, HLAB*35:01, HLA -C*01:05, and HLA-DPB1*11:01 were associated with susceptibility, although this was not statistically significant due to small cohort size. The authors concluded that human leukocyte antigen haplotype might affect immunity and promote an autoimmune process leading to the pathogenesis of nodding syndrome in the South Sudanese children.

Benedek and colleagues performed an immunogenetic analysis of macrophage migration inhibitory factor (MIF), a regulatory cytokine involved in a variety of pathologies such as epilepsy, neurodegenerative diseases, and autoimmune disorders, in 49 confirmed nodding syndrome cases (05). The cases were found to have significantly lower MIF-173C/G containing genotype compared to healthy controls. However, serum level of MIF was elevated in individuals with nodding syndrome compared to controls. The authors interpreted this as certain MIF genotypes are associated with disease protection, whereas elevated MIF levels in the plasma may lead to autoimmunity drive neuroinflammation. Larger immunogenetic studies, particularly among other affected populations, are needed to further validate the findings of this study.

Metabolic aspects. Inborn errors of metabolism, though infrequent in populations, may frequently result in epilepsy, especially in newborns and infants. However, there are no specific features of nodding syndrome that specifically point toward errors of metabolism. At present, two studies have described metabolic acidosis and diminished vitamin B6 levels in patients with nodding syndrome; further investigations are lacking (09; 13; 26; 53). Edridge and colleagues evaluated untargeted metabolomics in plasma and CSF and did not find any major differences between the cases and controls (12).

Pathology. Pollanen and colleagues reported neuropathological findings in five cases of nodding syndrome who were 13 to 18 years of age at death (47). Pathological examination showed tau-immunoreactive neuronal neurofibrillary tangles, pretangles, neuropil threads, and dot-like lesions affecting the cerebral cortex. The tau pathology preferentially affected the frontal and temporal lobes in a patchy distribution. In addition, subcortical nuclei and brainstem were also noted to be involved. The mesencephalopontine tegmental nuclei, substantia nigra, and locus coeruleus showed globes neurofibrillary tangles and thread (47). Pollanen and colleagues published a second study that assessed the neuropathologic features of 16 nodding syndrome cases in northern Uganda (46). The primary neuropathologic findings were tau pathology (16 out of 16), cerebellar degeneration (11 out of 16) and white matter degeneration (7 out of 16). The tau pathology was described as tau-positive deposits presenting as neurofibrillary tangles, pre-tangles and dot-like grains, and threads in neuropils. The regions commonly affected were the neocortex and locus coeruleus with involvement of the substantia nigra and tegmental nuclei.

In another independent study from Uganda, Hotterbeekx and colleagues described pathological changes in five individuals with nodding syndrome (15). The authors found inflammatory gliotic changes in several cortical regions and the mesencephalon. The authors also reported CD-68 positive macrophage cluster, primarily in the dentate nucleus, cerebellar vermis, medulla oblangata, and frontal cortex. Similar to the previously reported findings, Hotterbeekx and colleagues also noted presence of Tau-immunoreactive neurofibrillatory tangles in the cortical regions (temporal, frontal, and parietal lobes and the sulcus calcarinus), mesencephalon, thalamic, and basal nucleus.

Epidemiology

Population-based studies from three independent sites in sub-Saharan Africa--South Sudan, northern Uganda, and southern Tanzania--have been widely published (02; 23; 29; 63; 64; 65; 66; 27; 41; 58; 13; 18; 56; 57; 20). Elsewhere, van der Waals and colleagues reported a few cases of “seizure disorders with dorsoventral movements of the head” in Liberia, and Prischich and colleagues described four patients in Cameroon with complex partial seizures characterized by “repeated head movements” and impaired responsiveness lasting a few seconds to several minutes (62; 48). Additionally, a larger epidemiological study confirmed the findings of the smaller case series and found head nodding seizures in Cameroon (51). In Central African Republic, five cases of nodding syndrome were described in a cross-sectional survey of over 6000 individuals residing along the Ubangui river (37). Nodding syndrome appears to be “epidemic” in South Sudan and northern Uganda, although it seems rather endemic in Tanzania, with reports dating back to the 1960s (01; 02).

Uganda. In Uganda, the first cases of nodding syndrome were reported in 1997, and between 2000 and 2009 the numbers increased in an epidemic fashion, especially in the northern districts of Kitgum, Lamwo, and Pader (13). As of February 2012, the Ugandan Ministry of Health reported a total of 3097 suspected cases and 170 deaths (case fatality rate: 5.4%) from the three northern districts of Kitgum, Lamwo, and Pader; however, a decline in overall and newly diagnosed nodding syndrome cases was noted from mid-2012 onwards (44). Within the first 15 weeks of 2014, the Ministry of Health reported no new cases in Uganda. During the same period in 2013, 25 new cases were reported (60).

Using the newly developed consensus case definition, the U.S. Centers for Disease Control and Prevention and the Ugandan Ministry of Health performed a first systematic prevalence study in March 2013. In the three northern Uganda districts, 1687 probable nodding syndrome cases were registered, reflecting a prevalence of 6.8 probable nodding syndrome cases per 1000 children aged 5 to 18 years with a male to female ratio of 1.1 (20). It was noted that the cases were located in proximity to the fast-flowing rivers Pager and Aswa and breeding sites of Simulium flies; the term “river epilepsy” was suggested in accordance to “river blindness” (45; 59).

In northern Uganda, certain risk factors have been associated with the development of nodding syndrome. Gumisiriza and colleagues conducted a case-control study with a triad of 154 persons with nodding syndrome, persons with non-nodding syndrome epilepsies and healthy community controls (14). The participants were aged between 8 to 20 years and had their medical, sociodemographic, family, and migration histories reviewed. Seropositive O volvulus antibodies was identified in both persons with nodding syndrome and other non-nodding syndrome epilepsies. Having a preterm birth was suggested to be a potential cofactor after its association was found only in the nodding syndrome cohort. Living in an IDP camp was not associated with higher risk of nodding syndrome or other non-nodding syndrome epilepsies.

South Sudan. In Mundri County, in then southern Sudan, nodding syndrome emerged around 1991 and was reported around 1992 (57). This was prior to the large-scale displacement in the Jambo area noted in Lui, the later epicenter of nodding syndrome, around 1995 (29; 41). Increasing numbers were reported to the WHO in 1997, although incidence stabilized after 2000 with 300 cases being described by 2003 (28). Another increase in incidence of nodding syndrome was confirmed in 2010 (06). Most of the cases were reported from the town of Lui and the village of Amadi in Mundri County, Western Equatoria, with prevalence rates of 2.3% (41 of 1783) and 6.7% (57 of 854) in 2001, respectively, and a female to male ratio of 1.2 (58). The age at onset ranged from 2 to 18 years, with a mean duration of 3.0 years. Further cases were reported from Central Equatoria and Lakes State. In 2012, the overall number of cases was estimated by the Ministry of Health and WHO to be about 6000 to 8000 cases (52). As in Uganda, nodding syndrome is mainly seen in areas endemic for onchocerciasis and around the breeding sites of Simulium flies. It has only been reported in the sessile Moru tribe and not in the itinerant Dinka tribe (57). The Greater Mundri area remains a region of high burden for nodding syndrome cases. Abd-Elfarag and colleagues report a prevalence of 607 (2.7%) people with nodding syndrome out of 22,411 surveyed (03). One hundred and fourteen (19%) were new-onset cases, with the highest prevalence of nodding syndrome cases in the Diko village (13.7%). Nodding syndrome cases clustered around adjacent households and rivers. All behaviors around rivers (handwashing, bathing, drinking, cooking) and consumption of poultry were identified as potential risk factors for nodding syndrome. No relationship was found between internal displacement and prior ivermectin use.

Tanzania. Winkler and colleagues reported 62 patients with nodding syndrome with a median age of 14 years and a 3 to 2 female predominance (63). Detailed epidemiologic data from Tanzania are not available; so far, it has only been described in the Wapogoro tribe.

Differential diagnosis

The literature often links nodding syndrome to Nakalanga syndrome, a form of endemic dwarfism described in Uganda, Ethiopia, Burundi, and Rwanda (33; 21; 31; 25; 40). The children present with arrested growth in early or mid-childhood, delayed maturity and sexual development, wasting, muscle atrophy, and skeletal abnormalities (kyphoscoliosis, pectus carinatum, shortened neck, prognathism). The clinical findings have been attributed to endocrine dysfunction as histologic abnormalities in the adenohypophysis, thyroid, adrenal gland, and testis have been found. The children are often described as gentle, docile, and apathic; cognitive impairment and epilepsy are frequently associated, although not compulsory findings. The disorder was first described in areas hyperendemic for onchocerciasis, with a prevalence of up to 1.7 %. Virtually all patients suffered from onchocerciasis, and reports of the disorder ceased with eradication of Simulium vectors in the affected areas. Nevertheless, epileptic seizures seem to be only an additional finding in Nakalanga syndrome, although they are noted (21; 31; 25; 40). Stereotyped body movements such as nodding of the head or body rocking are observed in several neurologic and psychiatric disorders, such as Rett syndrome, autisms, and posttraumatic stress disorders. Many of the children, especially in Uganda and South Sudan are traumatized by displacement, violence, abuse and war crimes, orphanages, poverty, starvation, and lack of appropriate care, and their general appearance may resemble children from Romanian orphanages with anaclitic depression as reported in the ‘90s.

Diagnostic workup

At present, the diagnosis of nodding syndrome remains based on the epidemiological clinical case definition established at the First International Scientific Meeting on Nodding Syndrome and according to the level of evidence; suspected, probable, and confirmed cases are distinguished (67). There are no distinct serological, cerebrospinal fluid, imaging, or electrophysiological studies to diagnose nodding syndrome. However, nutritional and micronutrient deficiencies have been commonly reported in case of nodding syndrome and need to be evaluated. Iron deficiency, anemia, metabolic acidosis, and vitamin B6 deficiency are described, although they are likely secondary to the marked wasting. Kitara and colleagues described endocrine parameters such as thyroid-stimulating hormone, thyroxin (T4), T3, prolactin, vitamin D3, and parathyroid hormone in 10 patients without major abnormalities, but with no control group; the significance of these findings is unclear (26).

MRI studies performed by Idro and colleagues showed different degrees of cortical and cerebellar atrophy in 19 of 22 patients without focal cerebral cortical or hippocampal changes (17). Sejvar and colleagues performed MRI investigations in five children with nodding syndrome, four of whom showed varying degrees of generalized cortical and cerebellar atrophy most prominent in the parieto-occipital and anterior temporal areas and in the cerebellum (50). Hippocampal atrophy was noted in two children, and the MRI for one child revealed encephalomalacia with a large calcified cystic lesion of the right parietal lobe most likely due to a previous cerebral infection. Winkler and colleagues performed MRI studies in 12 patients (65). Four of seven patients with head nodding only had normal neuroimaging. In the other three patients, hippocampus pathologies were seen; two had additional gliotic lesions. None of the five patients with HN plus had a normal MRI; all showed either gliotic lesions or hippocampus pathologies.

In a study from Uganda, Mazumder and colleagues evaluated the structural MRI features of 39 nodding syndrome patients and compared their findings to age-matched controls with onchocerciasis-associated epilepsies (OAE) (36). MRI abnormalities were identified in 31 of 39 patients, with cerebral and cerebellar atrophy as the predominant characteristics in nodding syndrome and onchocerciasis-associated epilepsies patients.

Management

As the etiology and pathophysiology of nodding syndrome remains uncertain, there are no specific or definitive therapies available. Nevertheless, Idro and colleagues proposed guidelines for the management of nodding syndrome and presented a manual to guide and train health workers (17). Including data from Uganda, Tanzania, and South Sudan, they provide recommendations on symptomatic supportive care for emergency, inpatient, outpatient, and community-based settings and advise on diagnostics, follow-up care surveillance, documentation, and community education. Recommendations on seizure control, management of malnutrition, behavioral and psychiatric problems, nursing, wound care, infections, as well as nutritional, physical, and cognitive rehabilitation are given. Regarding seizure control, few data are available. Drugs were mainly chosen due to their availability, and, hence, their use differs at the various sites. At present, sodium valproate, phenobarbital, carbamazepine, and phenytoin have been described.

Within their guidelines, Idro and colleagues recommend valproate up to 40 mg/kg/day and suggest lamotrigine and levetiracetam as second-line medication, although both drugs have not been used in nodding syndrome so far (17). Recommendations on the duration of treatment have not been given.

Applying these guidelines in a cohort of 22 children previously treated with low doses of phenytoin, phenobarbital, and carbamazepine, Idro and colleagues reported a 57% reduction of seizure frequency (nodding and generalized tonic-clonic seizures) under valproate 15 to 25 mg/kg/day within 2 to 3 weeks (18).

Winkler and colleagues reported on the long-term outcome of 53 Tanzanian patients with nodding syndrome, 46 (86.8%) of whom were on antiepileptic medication (66). Of the 32 patients treated with phenobarbital, head nodding stopped in 15 patients, although 13 (86.7%) had ongoing generalized tonic-clonic seizures. Head nodding frequency decreased in 11 children, remained unchanged in four, and increased in two. In three of seven patients treated with carbamazepine alone, head nodding stopped, although the patients continued to have generalized tonic-clonic seizures. Head nodding seizures decreased in another three patients, whereas they increased in one patient.

With the findings of possible autoantibody mediated pathogenesis of nodding syndrome caused by molecular mimicry with O volvulus, patients were thought to benefit from immunomodulatory therapies (22). At the National Institute of Health, two cases were treated with five cycles of plasmapheresis administered every other day, and one case was treated empirically with intravenous immunoglobulins (2 mg/kg divided over two days). In the individuals treated with plasmapheresis, the serum levels of ov16 and Ov-tmy-1 specific antibodies decreased. The IVIG treatment did not have any effect on the serum levels of auto-antibodies. Clinical status and the EEG for none of the three cases treated with immunomodulation changed. The lack of response was thought due to long-standing epileptogenesis process established early during the onset of the disease, and the ongoing immune responses are not needed to maintain the disease process.