Eighty-three children presented at Goroka Base Hospital in the Eastern Highlands Province (EHP) of Papua New Guinea over a period of 3 years and 9 months between February 1997 and November 2000 were confirmed to have subacute sclerosing panencephalitis (SSPE). Confirmation of the diagnosis was based on the demonstration of high titres of measles antibodies in the cerebrospinal fluid and/or serum in association with clinical features supportive of SSPE, including characteristic electroencephalographic changes and amplification of measles virus genome by reverse transcriptase polymerase chain reaction in some cases. The mean cerebrospinal fluid and serum enzyme immunoassay antibody levels among the SSPE patients were 38 250 and 860 580, respectively. The mean age of onset of SSPE was 7.9 ± 2.6 years and ranged between 2 and 14 years. The overall male to female ratio was 1.2:1 and 1.4:1 for EHP.
Papua New Guinea (PNG), with a reported incidence of 56 cases per million population below 20 years in 1990 for four highlands and two coastal provinces (Lucas et al. 1992) and 98 per million population below 20 years in Eastern Highlands Province (EHP) during 1997 and 1998 (Takasu et al. unpublished observation), has the highest recorded incidence in the world. Generally, a relatively high incidence of SSPE would be expected in developing countries because of a high amount of circulating measles. Moreover, this high burden of measles in developing countries affects an appreciable proportion of very young children. However, as mass measles vaccination is effective in reducing the incidence of SSPE, one would expect this to decrease with the introduction of vaccination programs (Miller 1983; Anlar et al. 2001; Dyken 2001). This is not the case in the eastern highlands of PNG where the incidence has remained very high (Takasu et al. unpublished observation).
Before 1982 when routine measles vaccination was introduced, measles was not considered to be a major public health problem in PNG. As there were only a few reported cases of measles and the case fatality was low at that time, vaccination was considered not to be cost-effective (PNG Department of Public Health 1974; Coakley et al. 1991). However, this has no longer been the case since the early 1980s, because measles has increasingly become an important public health problem throughout the country. Outbreaks occur at regular three- to four-yearly intervals throughout the country, including in the EHP, and contribute to high morbidity and substantial mortality (Coakley et al. 1991; Mgone et al. 2000). Associated with this is the highest incidence rate of SSPE that has been observed.
Characteristically, clinical features of SSPE appear insidiously long after measles infection that may or may not be recalled. This, along with the fact that confirmation requires the demonstration of high titres of measles antibodies in the cerebrospinal fluid (a test which may not be readily available in many developing countries), makes definite diagnosis difficult in these settings (Anonymous 1990). Moreover, ancillary diagnostic facilities like electroencephalography (EEG) are also often unavailable. Therefore, more often than not, the diagnosis of SSPE in many developing countries has to be made on the clinical manifestations at presentation and during the follow-up.
In our study in the highlands of PNG, the diagnosis of SSPE was based on the presence of the characteristic clinical features and confirmed by the demonstration of high titres of measles antibodies in the cerebrospinal fluid and serum. In some cases the clinical presentations were supported by the characteristic EEG changes consisting of high-voltage complexes (Yalaz et al. 1988) and by amplification of measles virus (MV) genome.
Following informed consent, 95 children with the clinical diagnosis of SSPE presented at Goroka Base General Hospital (GBGH) over a period of 3 years and 9 months from February 1997 through November 2000 were enrolled for the study. The clinical diagnosis was based on an insidious onset of motor dysfunction (that included myoclonic jerks, frequent falls and changing gait), behavioural changes, progressive mental deterioration, speech impairment, rigidity, spasticity, autonomic disturbances and coma. A detailed clinical history was taken for all patients, and each underwent a neurological examination. The findings were recorded on a standardised data collection form. The information collected included personal details, history of measles infection and vaccination, age of onset of SSPE and its clinical manifestations.
Laboratory and clinical investigations
To confirm the diagnosis of SSPE, 2 ml of blood and 1 ml of the cerebrospinal fluid was collected for serology and nucleic acid analysis. Quantitative serum and CSF MV-specific IgG antibody estimation was performed using an enzyme immunoassay (EIA) kit (Denka Seiken Co., Ltd, Tokyo, Japan), as described elsewhere (Takasu et al . unpublished observation). Briefly, before performing the EIA, sera were diluted 1 in 200 and CSF 1 in 20 with the buffer solution provided. EIA values were determined using a standard curve drawn up by plotting optical densities (OD) against a series of dilutions of a pooled antibody-containing serum. The EIA value thus determined refers to the maximum dilution at which the specimen gives a positive result for the presence of antibody. The EIA value of the diluted specimens with a passive haemagglutination (PHA) titre of 32 was defined as 4.0. The PHA titre therefore corresponds to the EIA value of the diluted specimen multiplied by 8 (32 ÷ 4). The ratio of the OD of pooled human sera containing undetectable MV IgG antibody (EIA value = 0) to the OD of pooled human sera with low-titre MV IgG antibody (EIA value = 4.0) was <0.6. In the current study, the measles antibody levels are expressed as EIA values of the undiluted specimens.
Reverse transcriptase-polymerase reaction (RT-PCR) was performed with primers that amplify the hypervariable region of the N gene of the MV using CSF and peripheral blood mononuclear cell (PBMC) specimens from 19 patients. Details of methods are reported elsewhere (Miki et al. 2002).
EEG was performed in six patients using a two-channel EEG machine at the beginning of the study and in eight patients using a 10-channel machine in the later part of the study.
The diagnosis of SSPE was based on a combination of clinical, laboratory and EEG criteria as follows: (1) progressive mental or motor function deterioration associated with a positive history or presence of sudden jerky movements or falls with or without observable myoclonus or atonia; (2) CSF and serum EIA values of >2000 and >200 000, respectively; (3) periodic synchronous discharges (PSDs) in the EEG comprising bilateral and simultaneous periodic high-voltage slow wave complexes appearing every 4–10 s (Sharp et al. 1991; Dogulu et al. 1995; Yaqub 1996); and (4) positive amplification of a nucleotide sequence of MV with RT-PCR in PBMC or CSF specimens. Criteria 1 and 2 were mandatory whereas 3 and 4 were not as typical EEG changes may not be observable in some early and late cases (Doğulu et al. 1995), typical PSDs may be masked by atypical EEG changes (Sharp et al. 1991) and the MV genome is not always amplifiable by PCR in clinical specimens of SSPE patients. SSPE was ruled out when the CSF or serum EIA value was below 40 or 400, respectively. Therefore, patients were classified as definite cases of SSPE when in addition to the mentioned neurological features they had both CSF and serum EIA values of >2000 and >200 000, respectively, with or without typical EEG changes or positive PCR amplification of MV genome. Patients were classified as probable SSPE cases when they fulfilled criterion 1 and had either CSF or serum EIA values above 1000 or 100 000, respectively, or when clinical features and the course of the disease were consistent with SSPE, although their records were incomplete such that criterion 1 could not be fully satisfied, and both CSF and EIA values were above 1000 and 100 000, respectively.
During the study period, 95 children presenting with clinical features suggestive of SSPE were seen at GBGH in the EHP of PNG. The diagnosis of SSPE was confirmed in 61 (64%) of the cases and 22 cases (23%) were classified as probable SSPE. The majority of the children (57) were from within EHP and 15 from the neighbouring Simbu Province, three from Western Highlands Province and one each from Madang and Enga Provinces. In six children the province of origin was unknown. Among those classified as being probable cases of SSPE all had very high serum EIA measles antibody levels of greater than 100 000, which in our laboratory are characteristic of SSPE. They were classified so because they either lacked CSF results or some clinical data, but for all purposes they were grouped together and simply treated as SSPE cases. The overall mean CSF and serum measles antibody EIA values in the affected children were 38 250 (range 1900–130 400) and 860 580 (range 123 200–>2 000 000), respectively. The mean CSF measles antibody EIA values in the SSPE and probable SSPE patients were 38 380 and 37 740 and the corresponding serum values were 905 400 and 726 000, respectively.
Age and sex distribution
Among the affected children were 46 boys and 37 girls, the male:female ratio being 1.2:1 overall. Among them, 33 boys and 24 girls were from EHP, giving a male:female ratio of 1.4:1 in the province. The age of onset ranged from 2 to 14 years, mean 7.9 ± 2.6 years and median 7.5 years. Most children (74%) became ill after 5 years of age.
Past history of measles infection and vaccination
Twenty-eight children had had measles; another 28 had not had measles, 15 were uncertain. For 12 children no history was available, including six with missing data and another six from whom no such information was collected. Forty-five children has a positive history of measles vaccination with at least one dose of measles vaccine, but the dates of the immunisations were known for only 28 of them, including four children who did not have documentation to support this (Table 1). However, among the immunised children there were 13 (29%) who had suffered from measles, including nine who were vaccinated after and four before acquiring measles. Among the four children who were vaccinated before acquiring measles, two were vaccinated in less than 3 weeks before getting measles. In the remaining children, history of immunisation was uncertain in 18, unknown in 13 and denied in 7 children. The mean age of the first measles vaccination was 17.3 months (range 4 months–11.5 years) among all children with a positive history of immunisation and 18.7 months among those whose history was validated with documentation.
Table 1. Age at first measles vaccination and measles infection among the subacute sclerosing panencephalitis patients
Age in months
Positive measles vaccination
Positive history of past measles infection
The mean and median age of contracting measles infection among the 24 children who gave a history of having measles before developing SSPE and could recall at what age the infection occurred were 19.4 and 10 months, ranging between 2 months and 8 years. Among these, the information was verified in the child health record books in nine cases. In the cases verified, the mean age for contracting measles was 8.8 ± 2.7 months (range 6–14 months) and median was 9 months. The majority (63%) of the children who had measles contracted the infection in the first year of life as shown in Table 1.
The mean age at which SSPE was manifested was 7.9 ± 2.6 years (range 2–14 years and median 7.5 years). In the 24 children in whom a history of measles prior to the development of SSPE was recalled, it was possible to determine the time interval between measles attack and the development of the disease. In these the mean time interval between measles illness and onset of SSPE was 6.2 ± 1.9 years (range 2–11 years and median 6.2 years). There was no significant difference in the age of onset of SSPE among the children who had a positive past history of measles and those who did not.
Presenting clinical features
The information on clinical features among the SSPE patients was available in 80 patients though this was not always complete. Characteristically the onset of symptoms was insidious in nature and generally no specific precipitating factors or prodromal illness were recalled. However, 32% of the patients admitted to history of fever within 2 weeks prior to the onset of the symptoms. The mean duration of symptoms at the time of the diagnosis was 3.5 ± 4.0 months and ranged from 4 days to 3.5 years, pertaining mostly to central nervous system affliction as shown in Table 2. At the time of the diagnosis most patients presented with a disability, the majority of them in Stage 2 according to Jabbour's classification (Jabbour et al. 1969). Besides myoclonic jerks and falling attacks, which were part of the mandatory diagnostic criteria, common presenting features were changing gait, abnormal movements, speech impairment and inability to walk or stand.
Table 2. Clinical features of subacute sclerosing panencephalitis at the time of enrolment into the study
Inability to walk
Inability to stand
Loss of speech
Inability to feed
Impairment of consciousness
Electroencephalography was carried out in 14 children. Among these typical periodic synchronous discharge (shown in Figure 1) characteristic of SSPE was noted in four children.
Additionally, measles virus genome RT-PCR amplification was performed using CSF and peripheral lymphocytes in 19 SSPE patients. This was positive from peripheral lymphocytes of two patients and none from the CSF.
SSPE is a rare complication of measles that in many parts of the world has an annual incidence of only 1–3 cases per million population under the age of 20 years (Soffer et al. 1976; Bakir et al. 1988; Anonymous 1990; Sanders et al. 1990). In the current study conducted over a period of 3 years and 9 months, we report on 83 consecutive patients newly diagnosed to have SSPE. The diagnosis was based on clinical features and high titres of measles virus-specific antibodies in the cerebrospinal fluid and serum. Fifty-seven of these cases were from the EHP where a high incidence of SSPE has been previously reported (Sanders et al. 1990; Lucas et al. 1992) and recently confirmed to have remained high at an annual incidence of 98 per million population under 20 years for the period between 1997 and 1998 (Takasu et al. unpublished observation). Although geographical clustering of cases and ethnic difference in the incidence of SSPE has been reported before, this has not been fully explained (Jabbour et al. 1972; Mackenzie et al. 1975; Sanders et al. 1990). Similarly, genetic factors have also been implicated as a predisposing factor contributing to the high incidence of SSPE in the highlands of PNG (Lucas 1993) though there is no clear evidence to support this other than the fact that most of the cases came from the highlands.
Following the introduction of routine measles immunisation through the expanded programme for immunisation (EPI) in 1982 in PNG, the incidence of SSPE should have been expected to decline as has happened in many other countries including low-income ones (Miller 1983; Bloch et al. 1985; Vucenovic & Vranjesevic 1989). However, this has not happened. Among the possible causes may be inadequate vaccination coverage, untimely vaccination and poor cold chain system. As the mean age of the onset of SSPE in this study, which was conducted between 1997 and 1999, was around 8 years, it is important to critically look at the measles immunisation policy and coverage of the period around 1989–1991. At the introduction of mass measles immunisation through EPI in PNG in 1982, measles vaccine was at first given at the age of 9 months. However, since 1990 PNG has adopted a policy of giving two doses of measles vaccine, the first at 6 months and the second dose at 9 months of age and also vaccinating at every opportunity. Nevertheless, despite this approach the measles vaccination coverage had generally been unsatisfactory throughout the country including in the EHP where measles outbreaks have continued to occur at three- to four-yearly cycles (Mgone et al. 2000). Between 1989 and 1991 the measles coverage among infants below 1 year in EHP was between 55 and 60%, but declined to 19% in 1993 and 8% in 1994 (Papua New Guinea Department of Health 1986, 1998). During this period there have been two measles outbreaks, in 1992 and 1998, and just before that in 1986 and 1988 (Mgone et al. 2000). Associated with the low coverage is the late timing of measles vaccination in many children. In the current study, the mean age at acquiring measles infection among SSPE children with validated histories was around 9 months, whereas the mean age of measles vaccination was around 19 months. Normally regular attendance to well-baby clinics declines as the child gets older and quite often immunisations that are given in the later months of infancy are missed. Additionally, surveys that were conducted during this period by the Papua New Guinea Department of Health and the World Health Organisation had found serious problems and systems failure in the cold-chain facilities and immunisation delivery systems as well as frequent outbreaks of measles (Bass 1993).
The mean age of 7.9 years for the onset of the manifestation of SSPE in the current study is higher than the 4.9 years that has been reported previously from this country (Lucas et al. 1992). Similar increases in the age of onset in the manifestation of SSPE have been observed in other studies (Dyken 1985; Carman & Johnson 1989; Anonymous 1990). This has been attributed to the shift in age of contracting measles towards older children. However, in our study this shift in age of acquiring measles was not observed. A previous study conducted at the same hospital reported a mean age of measles attack of 17 months among measles admissions and 29 months in outpatient cases (Coakley et al. 1991). In a more recent study that was conducted during an outbreak of measles in the same hospital and around the same time when the SSPE cases were enrolled into this study, we found that the median age of measles illness among hospitalised children was 11 months and that 55% of the cases were infants under 1 year, 27% being under 6 months of age (Mgone et al. 2000). Early age of measles infection has been shown to be a major predisposing factor in the development of SSPE. However, this has been noted mainly from developed countries where measles generally attacks older children. As in developing countries measles infection is very common and quite often attacks very young children, this effect may be somewhat obscured. Moreover, contrary to the observation that early onset of measles predisposes to the development of SSPE studies from South Africa have shown that the age of contracting measles had no effect on the development of SSPE later in life (Schoub et al. 1990). Similarly, after correcting for infant mortality, Takasu et al. (1992) in Karachi found an elevated proportion of late measles among SSPE patients in comparison to early measles preponderance among children in the general population.
There are now several studies that contradict the male preponderance among SSPE patients that has been observed in the past (Bakir et al. 1988; Carman & Johnson 1989; Schoub et al. 1990). Moreover, some recent studies have shown not only a decrease in male preponderance but also a shift towards increased female preponderance especially among the older children (Modie et al. 1980; Carman & Johnson 1989). Unlike the previous study that was conducted in PNG which showed a slight male preponderance with a male:female ratio of 1.8:1 (Lucas et al. 1992), in the current study we have observed no male preponderance.
Although SSPE has been shown to account for a substantial proportion of childhood neourodegenerative conditions in developing countries (Kondo et al. 1988; Radhakrishnan et al. 1988; Yalaz et al. 1988), lack of diagnostic facilities means that most of the times clinicians have to make a diagnosis based on clinical features only (Anonymous 1990). Therefore, a good understanding of the clinical presentation of this condition as seen in developing countries is of a great importance.
In the current study the reported mean duration of symptoms at the time of presentation at the hospital was approximately 3 months, ranging from 4 days to 3.5 years. This relative lateness in reporting to the hospital may be due to a combination of factors including reluctance to report to hospital, slowness in the referral system and the insidious nature of the illness. Although in the majority of cases there was no recall of any particular precipitating factors prior to the onset of the illness, in almost one-third of the patients there was a history of fever within the previous 2 weeks of onset. However, this may be due to a recall bias and in any case more often than not parents are likely to complain of fever whenever their children are feeling unwell. Among the commonest clinical features noted in these patients were myoclonic jerks, frequent falling, change in gait, speech impairment and general weakness.
Because of its insidious onset, delayed manifestation after measles and prolonged course, SSPE is a slow virus disease. Kuru, another neurological disease of the Eastern Highlands, was also once classified as a slow virus disease but is now known to be caused by a prion. The earliest symptoms of kuru are headache and joint pain, which soon lead to unsteadiness in standing or walking and involuntary shaking. However, unlike SSPE, kuru affected only a small area of the Eastern Highlands, among the Fore people and their close neighbours. It affected mostly women of all ages and children; only 2% of the cases were adult males (Alpers & Gajdusek 1965). This was because it was transmitted through ceremonial mortuary feasts in which mainly women and children ate brains and other organs of their deceased relatives. With the cessation of this practice the disease is now disappearing and is seen only sporadically among elderly adults of both sexes who must have acquired the disease during their childhood in 1950s. Therefore kuru cannot be confused with SSPE.
Although the pathogenesis of SSPE is not fully explained, accumulated evidence suggests that the disease arises from the persistence of altered measles virus in the brain (Schneider-Schaulies & ter Meulen 1999; Schneider-Schaulies et al. 1999). However, even with the advent of highly sensitive technologies such as RT-PCR, measles virus has been successfully detected in very few clinical specimens from SSPE patients (Nakayama et al. 1995; Vardas et al. 1999). In the current study we were able to demonstrate the MV genome from only two cases among the 19 cases on which underwent the RT-PCR amplification (Miki et al. 2002).
Despite two decades of routine mass childhood measles vaccination in PNG, the incidence of SSPE remains high. This may be due to the inadequate vaccine coverage and logistics, as well as lack of timely measles immunisation owing to the underutilization of the existing services. However, as these constraints on immunisation programs apply universally to all low-income countries there might be other unexplained local factors that prevail in PNG. These may be environmental or genetic. Nevertheless, to reduce the current high incidence of SSPE in PNG, major improvements in measles immunisation coverage and the quality of the immunisation program must take place. This has been effective in the reduction of SSPE worldwide, including in other low-income countries.
We acknowledge the technical help by Mrs Hanayo Namae and Yumiko Saito, Department of Infection and Immunology, SRL, Tokyo, in performing the enzyme immunoassays and Mr Goni Merenge, EEG technician and nurse at GBGH, in recording EEGs and Dr Yuji Kokubun for training him. We thank all the children and parents, and all the PNG Institute of Medical Research and GBGH staff who have been involved with the study. This study was supported by funds from the Ministry of Education, Science and Culture of Japan, Nihon University, the Ministry of Health and Welfare of Japan and the Volunteers’ Organisation for the Elimination of SSPE in Goroka supported by the Ministry of Foreign Affairs of Japan Subsidy for Non-Governmental International Cooperation Activities through donation of EEG facilities and the training of Mr Merenge in EEG technology at Nihon University Itabashi Hospital, Tokyo, Japan.
Authors Prof. Michael P. Alpers, Papua New Guinea Institute of Medical Research, PO Box 60, Goroka EHP 441, Papua New Guinea. E-mail: M.Alpers@curtin.edu.au Peter G. Asuo, Goroka Base Hospital, PO Box 392, Goroka, EHP 441, Papua New Guinea. E-mail: firstname.lastname@example.org Ryuta Kawanishi, Department of Neurology, Nihon University School of Medicine, 30-1 Oyaguchikami-machi, Itabasi-ku, Tokyo, 173-8610, Japan. E-mail: email@example.com Katsuhiro Komase, The Kitasato Institute, Division of Research and Development, Research Centre for Biologicals, 5-9-1 Shirokane, Minato-ku, Tokyo 108-8642, Japan. E-mail: firstname.lastname@example.org Jacinta Kono, Papua New Guinea Institute of Medical Research, PO Box 60, Goroka EHP 441, Papua New Guinea. E-mail: email@example.com Prof. Charles S. Mgone, Papua New Guinea, Institute of Medical Research, PO Box 60, Goroka EHP 441, Papua New Guinea. Present address: The African Malaria Network Trust, PO Box 33207, Dar es Salaam, Tanzania. E-mail: firstname.lastname@example.org (corresponding author). Joyce M. Mgone, Goroka Base Hospital, PO Box 392, Goroka, EHP 441, Papua New Guinea. E-mail: email@example.com Kenji Miki, Department of Neurology, Nihon University School of Medicine, 30-1 Oyaguchikami-machi, Itabasi-ku, Tokyo, 173-8610, Japan. E-mail: firstname.lastname@example.org Prof. Toshiaki Takasu, Department of Neurology, Nihon University School of Medicine, 30-1 Oyaguchikami-machi, Itabasi-ku, Tokyo, 173-8610, Japan. E-mail: email@example.com