OUP user menu

A clinical and pathological study of motor neurone disease on Guam

H. R. Morris, S. Al-Sarraj, C. Schwab, K. Gwinn-Hardy, J. Perez-Tur, N. W. Wood, J. Hardy, A. J. Lees, P. L. McGeer, S. E. Daniel, J. C. Steele
DOI: http://dx.doi.org/10.1093/brain/124.11.2215 2215-2222 First published online: 1 November 2001


Despite over 40 years of intensive study, the cause of the high incidence of motor neurone disease (MND) on Guam, and the relationship between this disease and MND seen in the rest of the world are still uncertain. We present a series of 45 cases of Guamanian MND, which reaffirm the clinical similarity between this disease and MND seen in other countries. However, the occurrence of MND among the indigenous Chamorros of Guam is distinguished by four factors: (i) high prevalence; (ii) frequent familial occurrence; (iii) co-occurrence with the parkinsonism–dementia complex (PDC); and (iv) association with an unusual and distinctive linear retinopathy termed Guam retinal pigment epitheliopathy (GRPE). These distinguishing factors were not present in four non-Chamorros who resided on Guam when their MND symptoms occurred. Pathologically, the classical features of MND were seen in Guamanian Chamorro cases including ubiquitin inclusions. Neurofibrillary tangles were frequently seen. The neurofibrillary tangles appeared in the same distribution as described in the PDC but, unlike classical PDC, they were not usually associated with cell loss and occurred less frequently. While neurofibrillary tangle formation and the clinicopathological syndrome of MND may occur in parallel, observations from this series suggest that pathologically classical MND on Guam may occur independently of neurofibrillary degeneration and the clinical features of PDC.

  • Guam amyotrophic lateral sclerosis
  • motor neurone disease
  • neurofibrillary tangles
  • parkinsonism-dementia complex
  • pathology
  • ALS = amyotrophic lateral sclerosis
  • GRPE = Guam retinal pigment epitheliopathy
  • MND = motor neurone disease
  • PDC = parkinsonism–dementia complex


The description of a high prevalence focus of amyotrophic lateral sclerosis (ALS) on the Western Pacific island of Guam by Kurland and Mulder (Kurland and Mulder, 1954; Mulder et al., 1954) has been followed by over 40 years of research in an attempt to define this disease. However, despite this thorough work, the three central questions raised by the high prevalence of ALS, or more broadly motor neurone disease (MND), on Guam remain unanswered: its aetiology, its relationship to parkinsonism–dementia complex (PDC) and its relationship to classical ALS. The earliest reported diagnosis of progressive muscular atrophy on Guam was made in 1904, and ALS was subsequently reported at a high incidence on Guam by Zimmerman and others (Zimmerman, 1945). The initial surveys of Kurland and Mulder found that the clinical features of ALS on Guam were identical to classical ALS, aside from the strong family history of ALS in the Guamanian patients (Kurland and Mulder, 1954; Mulder et al., 1954). Zimmerman's pathological findings and Mulder's clinical opinion that ALS on Guam was the same disease as that seen in Europe and America seemed less likely when Hirano and colleagues published their clinical and pathological description of PDC (Hirano et al., 1961a, 1961b). They described an unusual pattern of neurofibrillary degeneration affecting the mesial temporal cortex, basal ganglia and brainstem, and designated this syndrome the parkinsonism–dementia complex. They described an identical topographical pathology in Guamanian patients with ALS, and concluded that the two clinical syndromes were different manifestations of a single disease and represented two ends of a unitary clinical spectrum (Hirano et al., 1961a, 1961b). In 1994, Oyanagi and colleagues re-evaluated the pathology of ALS on Guam using ubiquitin and tau immunohistochemistry (Oyanagi et al., 1994). They described the pathology of classical ALS, including ubiquitin deposits, in the anterior horn cells of patients with Guamanian ALS, which seemed to occur independently of the deposition of tau-containing neurofibrillary tangles. In contrast to Hirano, Oyanagi concluded that PDC and ALS were separate diseases and that ALS on Guam is classical ALS and identical to ALS in America and Europe. This recent suggestion that ALS and PDC are separate diseases is further supported by epidemiological trends which show that the incidences of ALS and PDC are changing independently, with the incidence of ALS declining very rapidly in recent years (Waring, 1994). The clinical profile of ALS/MND on Guam has not been considered since a review of the 30-year experience of both PDC and ALS, from 1950 to 1979, published in 1986 (Rodgers-Johnson et al., 1986).

This paper presents a retrospective review of 45 cases of MND seen on Guam between 1983 and 1998, and considers their clinical and pathological similarity to classical MND elsewhere in the world.


The clinical features of every case of MND encountered by one of us (J.C.S.) during 15 years of neurological practice (1983–98) on Guam were reviewed. These patients were recruited via referral from primary and hospital physicians and house-to-house surveys in the southern villages of Guam. Follow-up was made by clinical review, telephone contact and interviews with relatives for patients who had left Guam; one patient was lost to follow-up.

Following informed consent, pathological confirmation of diagnosis was established for 16 cases and brief details of these pathological examinations are presented. In 11 cases archival material was available for review. These cases were processed, sectioned and stained using routine histological stains and examined immunohistochemically using antibodies to tau (AT8, monoclonal mouse 1 : 200; Innogenetics, Zwijnaarde, Belgium) and ubiquitin (ubiquitin, polyclonal rabbit, 1 : 500; Dako Ltd, Ely, UK). Where available, hippocampus and spinal cord were studied along with motor cortex, medulla and cerebellum. In cases or individual case areas in which archival material was not available for review, a pathological summary was produced based on the contemporaneous autopsy report using routine histological stains, usually including silver staining but not immunohistochemistry.


Clinical features (Table 1)

View this table:
Table 1

Clinical features

nM : FMean onset age (years)Mean survival in years (range)Median survival (years)Pathology confirmedGRPE (of 19 tested)Hyposmia (of 13 tested)EPSDemFH ALS aloneFH ALS +PDCFH PDC alone
M = male; F = female; GRPE = Guam retinal pigment epitheliopathy; EPS = extrapyramidal syndrome; FH = family history; Dem = dementia; PMA = progressive muscular atrophy. *Cases presenting between 1980 and 1989.
Total451.8 : 149.4 9.4 (1–41)5.035.0%47.0%53.0%13.3%8.8%22.2%15.0%6.7%
1980–1989 cohort*2153.6 4.53.0
ALS3949.1 9.64.5
PMA353.3 3.32.0
Suspected ALS349.313.76.0
Chamorro4148.6 9.85.534.0%52.6%14.6%9.7%24.0%17.0%7.3%
Non-Chamorro458.0 6.01.550.0%0.0%0.0%0.0%0.0%0.0%0.0%
Bulbar onset755.6 3.32.0
Limb onset3838.310.66.0

Forty-five patients (41 Chamorros and four non-Chamorros) were diagnosed as having MNDs. Of these, 37 met the El Escorial criteria (Brooks, 1994) for clinically definite or probable ALS, three had a progressive muscular atrophy type syndrome and three had a diagnosis of possible ALS with a non-compressive monomelic mixed upper and lower motor neurone syndrome, or mixed upper and lower motor neurone signs without rostral corticobulbar/corticospinal involvement. These cases of MND are described together (Table 1). For calculation of mean and median survival, a 10-year cohort of 21 patients with disease onset between 1980 and 1989 was used; one patient has been lost to follow-up. Only Chamorro patients had the following features: a mixed syndrome with extrapyramidal signs and dementia, a positive family history of both ALS and PDC and Guam retinal pigment epitheliopathy (GRPE) (although only one non-Chamorro was assessed for that condition). Patients with ALS are more likely to have a family history of ALS in at least one first degree relative than PDC (37.2 versus 26.1%). In six patients with concurrent extrapyramidal signs, three presented with limb or bulbar weakness, one presented with memory impairment and two presented with general slowness. Three of these patients had tremor, cogwheel rigidity and bradykinesia, and in one of these patients the extrapyramidal signs were asymmetrical. A history of significant residence away from Guam is available for five patients who left Guam on average at 20 years of age (range 18–24 years), returned at the age of 41 years (range 37–42 years) and developed the disease at the age of 52 years (range 37–66 years). One case developed symptoms whilst resident overseas. The four non-Chamorro cases were two Caucasians and two Filipinos who immigrated to Guam at the ages of 32, 39, 43 and 29 years (in 1955in 1971in 1969 and 1970) and developed symptoms at the age of 63, 60, 52 and 57 years, respectively (mean duration from immigration to disease onset 20.25 years with a range of 21–28 years).


The classical features of ALS were present in the 16 cases examined (Table 2). These features included anterior horn cell loss and corticospinal tract degeneration. The hippocampi of nine cases were available for immunohistochemical examination. None had ubiquitin-positive inclusions of the type that may be seen in MND or MND inclusion dementia. Six (67%) had moderately or markedly severe tau positive neurofibrillary tangle formation, more marked than that seen in Caucasian individuals of a similar age. The average age of death of individuals with immunohistochemically defined moderately or markedly severe hippocampal neurofibrillary tangle formation was 54 years. However, this neurofibrillary tangle formation was not associated with significant nerve cell loss, and this included Case 9 who had clinically diagnosed mixed ALS/PDC. Two of the cases with markedly severe hippocampal neurofibrillary tangle formation had concurrent senile plaques.

View this table:
Table 2

Pathological features of ALS on Guam

CaseDiagnosisEthnicityAge at death (years)Survival (years)Tau +ve hippocampal NFTUb +ve hippocampal inclusionsHippocampal cell lossTau +ve frontal cortex NFTSubstantia nigra cell lossSPCorticospinal tract degenerationTau +ve AHC inclusions Ubiqitin +ve AHC inclusionsAHC cell loss
– = absent; +/– = mild; + = present; ++ = moderately severe; +++ = markedly severe; n.a. = not available; Cham = Chamorro; Cauc = Caucasian; NFT = only evaluated by silver or haematoxylin and eosin staining. *Archival material not available for immunohistochemical examination.

Spinal cord was available for re-examination in 10 cases. Six of these cases had ubiquitin-positive tau-negative inclusions, with a filamentous skein-like morphology typical of ALS/MND (Figs 1 and 2). Only one case out of 10 was identified which had tau-positive ubiquitin-negative inclusions in anterior horn cells.

Fig. 1

Ubiquitin immunohistochemistry of cervical anterior horn cell, showing skein-like inclusion. Scale bar = 125 μm.

Fig. 2

Further examples of skein like inclusions (arrowed) in anterior horn cells, ubiquitin immunohisto-chemistry. Scale bar = 30 μM.


This series provides data on patients with MNDs who developed symptoms on Guam from 1951 onwards, and who have been followed-up to 1998. Although most research work on Guam has concentrated on the possible aetiologies of the Guamanian neurodegenerative diseases, an understanding of the clinical features and the clinicopathological correlation is an essential background to this investigation. Whether there are one or two neurodegenerative diseases on Guam has important implications for interpreting the pattern of the familial occurrence of these diseases and understanding the epidemiology. The observation of neurofibrillary tangles in ALS has led to the view that there is one disease on Guam with varied clinical manifestations, and that it represents a unique form of ALS known as `Western Pacific ALS' (Hirano et al., 1961b). Our clinical and pathological observations, together with those of Oyanagi and colleagues, suggest that this may not be the case, although this depends on the background level of neurofibrillary tangle formation among asymptomatic Chamorro subjects.

The clinical features of the majority of patients in our series meet the El Escorial criteria for the diagnosis of ALS (Brooks, 1994) with progressive bulbar and limb upper and lower motor neurone involvement without sensory signs or evidence of compressive pathology. The only exception is the presence of extrapyramidal features, which would be an exclusion criterion, in 13% of cases. In addition, we have identified patients with a progressive muscular atrophy phenotype. The average survival of the whole group is longer than is normally seen in ALS, but this is biased by the inclusion of long-lived patients whose disease onset occurred before 1981 and the exclusion of short-lived patients from that earlier period. Analysis of a 10-year cohort, with disease onset between 1980 and 1989, shows a mean and median survival of survival of 4.5 and 3 years, which is comparable with that seen in classical ALS (Alcaz et al., 1996). Similarly, the average age of onset of 49.4 years is similar to classical ALS, as is the reduced survival in patients with bulbar symptoms at onset. Although a survival time of one individual, with typical ALS, of 41 years is striking, long survivors with ALS have been described in non-Guamanian series, with 20% surviving for over 5 years in one study (Mulder and Howard, 1976). Overall, these findings do not support the view that ALS on Guam is more benign than that seen elsewhere in the world.

Four main features distinguish ALS on Guam: the high prevalence, the positive family history, the association with GRPE and the co-occurrence of parkinsonism and dementia. A frequent positive family history was noted in the early investigation of this disease, and it was assumed that the disease was most likely to be inherited (Mulder et al., 1954). Our series has confirmed that this familial predisposition continues, with a first degree family history of ALS in 37.2% of patients in this series. Although autosomal dominant ALS is well recognized, it accounts for only ~5% of all cases found outside of Guam (Mulder et al., 1986). The most common gene identified to be abnormal in familial ALS is SOD-1 (Cu/Zn superoxide dismutase), but analysis of this gene (Rosen et al., 1993) in eight patients with ALS on Guam has not revealed a mutation in affected subjects (Figlewicz et al., 1994). A genetic effect on the aetiology of these diseases remains uncertain, although a recent analysis has suggested that ALS and PDC may be due to the additive effect of two alleles (Bailey-Wilson et al., 1993). It is of interest that our data show that there is an excess occurrence of a family history of ALS as opposed to PDC in a ratio of 1.7 : 1 in these MND cases. This is seen despite the fact that during 1986–1987 the prevalence of PDC was estimated to be double that of ALS (Lavine et al., 1991). This can be compared with the segregation ratios in our previously reported series of autopsy cases of PDC (McGeer et al., 1997), in which we found that of 73 true siblings of the PDC cases, only eight (11%) were diagnosed with ALS, while 33 (45%) were already diagnosed with PDC, which is close to the 50% expected for autosomal dominant inheritance. Thus, PDC and ALS may cluster separately within either ALS or PDC families, as well as in families in which both diseases occur. We interpret these data as being consistent with the ALS and PDC syndromes of Guam being separate diseases. The common finding of a positive family history in patients with ALS, confirmed in a case control study (Plato et al., 1986), suggests that a familial habit, vertically transmitted infection or genetic factor may play an important role in the development of the disease. Conversely, the residential history in four cases in this series suggests that ALS also occurs in Guamanians who have not had a lifelong exposure to the Guamanian environment.

GRPE was first described by Cox who noted this condition in the course of investigating the supranuclear gaze palsy of PDC (Cox et al., 1989). This retinopathy has been demonstrated at an increased incidence among patients with ALS and/or PDC (~53%), at a lower incidence in asymptomatic Guamanian Chamorros and infrequently in Chamorros residing on other islands of the Mariana group, and thus mirrors the geographical distribution of both PDC and ALS (Cox et al., 1989). We have confirmed an increased incidence of GRPE in patients with ALS. Although the presence of GRPE is clearly not a prerequisite for the disease, it may be a marker, perhaps of infection with a parasitic agent which may be linked to the disease. Of note, GRPE is also seen in a large percentage of those Chamorros with parkinsonism and/or dementia (Cox et al., 1989).

Parkinsonism and dementia are present in 13% of our series of patients with ALS and it could be argued that this indicates that these syndromes are part of a single disease process. However, our family data together with that of other groups (Plato et al., 1986), indicate that these diseases can cluster together in different individuals in the same families. It is possible that PDC and ALS are separate diseases with a common genetic or environmental predisposition, which often occur together within the same families and occasionally occur together in the same individual. It is clear that dementia and parkinsonism are not invariable or even common accompaniments of ALS, and that many long-term survivors with ALS have not developed these clinical features. The explanation that the presenting clinical syndrome is determined by different rates of degeneration in different parts of the nervous system with all parts finally becoming affected, seems unlikely based on this evidence. However, it remains possible that these two clinicopathological phenotypes may reflect a regional variability in disease expression, with anterior horn cells producing ubiquitinated inclusions and hippocampal neurones producing tau neurofibrillary tangles in response to a single disease-provoking agent. We believe that the further clarification of the nosology of these diseases is most likely to occur at the pathological or genetic level.

Our pathological data confirm the long-standing finding of tau neurofibrillary degeneration in cases of ALS of Guam, and this occurs to a level which is not seen in asymptomatic Caucasian individuals of a similar age. However, in these cases, this neurofibrillary degeneration was not associated with significant nerve cell loss, clinical dementia or extrapyramidal syndromes. This is in contrast to the features of typical PDC, in which neurofibrillary degeneration is associated with substantial depletion of neurones in the temporal lobe, the CA1 region of the hippocampus, substantia nigra and brainstem nuclei, and significant dementia and movement disorders (Hirano et al., 1961b). It has been suggested that asymptomatic control Chamorro individuals have a higher level of neurofibrillary tangle formation without cell loss, and this phenomenon is positively associated with a family history of PDC (Anderson et al., 1979; Perl et al., 1995). It may be that a small amount of neurofibrillary tangle formation without senile plaque formation is an accompaniment of ageing (Delacourte et al., 1999), and it has been suggested that this process may be accelerated on Guam (Anderson et al., 1979). The presence of hippocampal neurofibrillary tangles may therefore be a background accompaniment to other neuropathological processes (Oyanagi et al., 1994). As our series does not include control cases, it is difficult to be certain whether the neurofibrillary tangle formation seen in these ALS cases is similar to asymptomatic Chamorro control levels. This can only be accurately evaluated with a prospective case control series.

One case out of 10 in this series has been identified to have tau-containing inclusions in anterior horn cells, as compared with six cases out of 10 with ubiquitin inclusions. This is of particular interest since some families with frontotemporal dementia linked to chromosome 17 such as a dementia disinhibition parkinsonism amyotrophy complex, resulting from a primary genetic abnormality of tau, may have a clinical anterior horn cell syndrome (Lynch et al., 1994). Furthermore, recently described transgenic mice which contain the tau P301L mutation have prominent clinical and pathological involvement of anterior horn cells with tau-containing anterior horn cell inclusions (Lewis et al., 2000). This suggests that primary tau disorders may produce both anterior horn cell and extrapyramidal/dementia syndromes. We have identified anterior horn cell tau inclusions in only one Guamanian case with MND, as compared with six cases with ubiquitin immunoreactive anterior horn cell inclusions. However, although these data are compatible with normal ALS type pathology in some—but not all—cases, the ability to define the typical ubiquitinated anterior horn cell inclusions may depend on the number of spinal cord levels studied, and in some cases the full spinal cord was not available. Overall, these anterior horn cell inclusions which are typically seen in classical MND (Figs 1 and 2) (Matsumoto et al., 1990b) support a classical MND process on Guam rather than a tau neurofibrillary tangle-related anterior horn cell disease. Previous studies of tau deposition in the spinal cord of patients with Guamanian MND have shown neurofibrillary tangle formation to be most marked in the posterior horn (Matsumoto et al., 1990a), with the anterior horn cells relatively spared. Conversely, dementia is well recognized in patients with MND, and despite the El Escorial criteria there are some reports of extrapyramidal features in patients with clinically and pathologically diagnosed MND (Desai and Swash, 1999; Neary et al., 2000). Some patients with frontotemporal dementia have MND-type inclusions in the hippocampus (`MND-inclusion dementia'; Jackson et al., 1996; Rossor et al., 2000), and the absence of ubiquitinated hippocampal inclusion in the Guamanian MND patients in this study suggests that the Guam disease does not overlap with MND inclusion frontotemporal dementia.

As previously reported, MND predominantly affects the Chamorro as opposed to the Filipino or Caucasian populations on Guam (~70 000 out of a total island population of 130 000). Only four MND cases were encountered in the non-Chamorro population over this 15-year series, and this does not support a widespread recent environmental exposure.

In recent years, the neuropathological description of PDC has become increasingly refined, with the description of selective cortical laminar involvement, specific tau protein deposition and of a unique tau glial pathology in PDC (Hof et al., 1994a, 1994b; Oyanagi et al., 1997; Perez-Tur et al., 1999). Some pathological issues are uncertain and need to be resolved to allow progression in the investigation of the disease, in particular the presence of neurofibrillary degeneration in non-Chamorro patients and its similarity to PDC, and the possible continuing presence of neurofibrillary tangles in asymptomatic younger Chamorros.

The clinical and pathological experience documented in this study leads us to the conclusion that ALS and PDC may be separate diseases which may be predisposed to by a single, or related, factor(s) on Guam. Despite the presence of neurofibrillary tangles suggestive of early PDC in some patients with ALS, we believe that the frequent lack of overlap in clinical syndromes and pathology, even in long survivors with ALS, suggests that these may be separate diseases.


We wish to thank Dr D. Mulder for helpful comments on the article and Drs C. Bergeron, A. Hirano, P. Lantos, J. Parisi and D. Perl for allowing us to include some of their pathological data. We wish to thank Mavis Kibble for technical assistance. H.R.M. is an MRC clinical training fellow and this work is supported by the Progressive Supranuclear Palsy (Europe) Association, the Mayo Foundation and the Society for Progressive Supranuclear Palsy.


View Abstract