American Journal of Medical Genetics (Neuropsychiatric Genetics) 88:4–10 (1999) Genetic Epidemiological Study of Schizophrenia in Palau, Micronesia: Prevalence and Familiality Marina Myles-Worsley,1* Hilary Coon,1 Josepha Tiobech,2 Jay Collier,2 Paul Dale,3 Paul Wender,1 Fred Reimherr,1 Anthony Polloi,2 and William Byerley1 1 Department of Psychiatry, University of Utah School of Medicine, Salt Lake City, Utah Belau National Hospital, Koror, Palau, Micronesia 3 University of Hawaii at Manoa, Honolulu, Hawaii 2 We are studying the genetic etiology of schizophrenia in the Republic of Palau, a remote island nation in Micronesia that has been geographically and ethnically isolated for approximately 2,000 years. The first epidemiological phase sought to estimate the lifetime prevalence of schizophrenia and evaluate the familiality of the illness based on complete ascertainment of cases and families segregating schizophrenia. A total of 160 strictly defined cases of schizophrenia were ascertained in a population of 13,750 adults who were 15 years of age and older. The lifetime prevalence of strictly defined schizophrenia in Palau was 1.99% overall and 2.77% in males vs. 1.24% in females. This greater than 2:1 male-to-female risk ratio for schizophrenia was accompanied by an earlier mean age of onset for males (23.3 years) than for females (27.5 years). These 160 cases of strict schizophrenia represent 59 separate families each identified by a single common founder. Eleven of these families have 5 to 14 cases and represent nearly half of the strict schizophrenia cases in Palau. Although schizophrenia is clearly aggregating in these 11 families, cases are distributed sparsely throughout the large sibships. In the entire sample of 160 cases of strict schizophrenia, there were only 11 sib-pairs and 2 sib-trios. When a family was defined to include third-degree relatives, only 11 cases (6.9%) were nonfamilial. The majority of the ascertained cases can be linked together into extended pedigrees with complex mul- Contract grant sponsor: National Institute of Mental Health; Contract grant numbers: MH54186, MH56098, MH10889, and MH52055. *Correspondence to: Marina Myles-Worsley, Department of Psychiatry, University of Utah School of Medicine, 50 North Medical Drive, Salt Lake City, UT 84132 Received 9 December 1997; Revised 11 March 1998 © 1999 Wiley-Liss, Inc. tilineal inheritance patterns. These intricately interconnected families may pose challenges for traditional linkage techniques. However, these Palauan families represent a valuable resource for studying the genetic etiology of schizophrenia because there may be fewer susceptibility genes for schizophrenia in this genetic isolate than in the heterogeneous populations that are common throughout the world today. Am. J. Med. Genet. (Neuropsychiatr. Genet.) 88:4–10, 1999. © 1999 Wiley-Liss, Inc. KEY WORDS: genetic isolate; multilineal pedigrees; gender differences; age-of-onset INTRODUCTION The importance of genetic factors in determining liability to schizophrenia has been firmly established by family, twin, and adoption studies [Gottesman et al., 1982, 1987]. Additional evidence for the genetic transmission of schizophrenia is provided by epidemiological studies that have found a remarkably narrow range of prevalence rates across populations living in vastly different geographic and socioeconomic environments. Estimates of lifetime risk for schizophrenia based on the World Health Organization Ten Country Study [Jablensky et al., 1992] range from 0.5% in Honolulu, Hawaii to 1.7% in Chandigarh, a rural area in India. However, pockets of elevated prevalence in the 2% to 3% range have been reported in certain regions of the world where geographical or cultural influences have resulted in genetically isolated populations. For example, lifetime morbid risk for schizophrenia is 2.4% in an isolated region in Northern Sweden [Book et al., 1978] and 3.2% in an internal isolate in Finland [Hovatta et al., 1997]. Genetically isolated populations like these represent a potentially valuable resource for studying the genetic etiology of schizophrenia because they may have fewer genes predisposing to schizophrenia than the genetically heterogeneous populations that are common throughout the world today. Schizophrenia in Palau, Micronesia Another genetic isolate where the rate of schizophrenia may show signs of elevation is Palau, Micronesia. Dale  studied the prevalence of schizophrenia throughout Micronesia and found evidence for a westto-east gradient with the highest rates in Palau and Yap to the west and the lowest rates in the easternmost island territories. The Republic of Palau, also known as Belau, is the westernmost archipelago in Micronesia. The Palau islands were probably settled by migrants from the Philippines and Indonesia approximately 2,000 years ago [Parmentier, 1987]. Archaeological studies report radiocarbon dates of 40 ± 70 AD for pottery fragments found in Palau [Takayama, 1981]. Blood group clustering, linguistic analysis, and ethnographic studies indicate that the Palauan population has developed in relative isolation, even from other Micronesian populations, since its original settlement [Simmons et al., 1965]. Over the past century, Palau has been ruled by Spain (1885–1898), Germany (1899– 1914), Japan (1914–1945), and finally the United States, which administered Palau as a United Nations Trusteeship from the time World War II ended in 1945 until 1994 when Palau became an independent nation. The Palauan population was reduced to 3,500 during World War II. Today, there are 20,470 native Palauans, 13,750 of whom are 15 years of age or older. Palau consists of more than 200 islands that stretch 125 miles along a northeast-to-southwest axis, but only five islands are inhabited. Over half the population resides on the island of Koror where the Belau National Hospital is situated, and the remaining Palauans live either on the adjoining island of Babeldaop, accessible by bridge, or on neighboring islands within a 1 h boat ride or 20 min plane trip from Koror. The basic unit of social organization in Palau is the clan, a blood-based, landholding group of families that share a common female founder. Ownership of land is inherited matrilineally, and every clan birth is on file at the Ministry of Justice. Clan members are expected to participate actively and financially in the traditional ceremonies that mark major life events such as births, marriages, and deaths. These communal events serve to teach younger members of the clan about their genealogy and clarify the blood relationships that bind the clan. Matings between clan members, even third or fourth degree relatives, have traditionally been forbidden. Consequently, inbreeding is rare in the Palauan population despite its isolation. Furthermore, outbreeding via marriage to outsiders has been rare because of the financial rights that accrue to native Palauans. Some Palauans have Asian or European ancestors, but such cases are the exception: the majority can trace their lineage back through five or more generations of full-blooded Palauans. There are none of the technical, linguistic, or practical barriers to conducting research that are usually found in remote, isolated locations. Since the country became a U.S. Trust Territory after World War II, Palau has evolved into a modern society. Both English and Palauan are official languages in Palau. The literacy rate is currently over 90%, and all Palauans who have been educated in Palau since the late 1940s have 5 received their elementary school through high school education in English. As a former U.S. Trust Territory, Palau’s medical services and research have been under the jurisdiction of the U.S. Public Health Service. All inpatient and outpatient services are administered by Belau National Hospital, and complete medical records written in English have been kept since the early 1950s. The first phase of our study examined the epidemiology of schizophrenia in Palau. Given the size of the population, the quality of the medical records at Belau National Hospital, and the spirit of co-operation of the Palauan people, complete ascertainment was considered feasible. Our aim was to diagnostically assess all cases of schizophrenia in Palau and construct a pedigree for each confirmed case by identifying all first, second, and third degree relatives. Based on these data, our objectives were to a) estimate lifetime morbid risk for schizophrenia in Palau including gender and age-of-onset characteristics and b) evaluate the familiality of the illness in this genetic isolate. MATERIALS AND METHODS Ascertainment of Cases and Families We used medical records on file at Belau National Hospital where all mental illness is treated to identify every Palauan who has ever been treated for a psychotic illness. For patients meeting our diagnostic criteria, we ascertained the individual’s first, second, and third degree relatives to identify other potentially affected family members and to reveal any relationships to previously ascertained families. Families were extended by identifying all offspring of the Generation I ‘‘founder’’ parents and ascertaining all subsequent descendants 15 years of age or older. The initial framework for family ascertainment was provided by pedigree diagrams drawn by Dr. Anthony Polloi, who returned from medical training in Fiji in 1968 to become Director of Mental Health at Belau National Hospital. Noting that many of the schizophrenic patients seeking regular treatment were members of several large clans, Dr. Polloi began constructing pedigrees and inserting them in medical records as part of the patient’s history. Psychiatric Assessment All living and consenting subjects were interviewed by Dr. William Byerley by using a modified version of the Schedule for Affective Disorders and Schizophrenia—Lifetime Version [SADS-L, Endicott and Spitzer, 1978]. Questions were added pertaining to childhood development and educational, occupational, adult medical, and female reproductive history. Age of onset was defined as the earliest age at which the presence of a clear psychotic symptom could be identified. All interviews were conducted in the presence of a Belau National Hospital Behavioral Health professional, primarily Josepha Tiobech, the Psychiatric Nurse Practitioner who has supervised psychiatric care in Palau for the past 10 years. When subjects preferred to be interviewed in their native language, each of Dr. Byerley’s questions was immediately repeated in Palauan, and the subject’s response was then translated into English 6 Myles-Worsley et al. and recorded by Dr. Byerley. For subjects who were deceased, we interviewed relatives with knowledge of the patient’s illness. The SADS-L interviews were supplemented by a review of psychiatric medical records. These records, which date back to the early 1960s and are written in English, document every inpatient and outpatient contact, which typically included presenting symptoms or complaints, summary of patient interview, diagnosis, and treatment plan. In addition, a psychosocial history was prepared by a Palauan social worker for every psychiatric patient receiving regular treatment. All available information on each subject was compiled with names deleted and identifying information such as race, religion, or unique cultural references generalized or disguised. The information was forwarded to two board certified psychiatrists (F.R. and P.H.W.) who have been conducting blind reviews of psychiatric patient data for ongoing genetic studies of schizophrenia and bipolar affective disorder at the University of Utah since 1986. After reviewing this information individually, Dr. Reimherr and Dr. Wender met and established consensual diagnoses according to RDC criteria [Spitzer et al., 1978]. If diagnostic disagreement occurred, an attempt was made to collect additional information for review. If agreement could not be reached, the subject was coded unknown. The RDC diagnoses included in the strict diagnostic category versus other psychotic illnesses are presented in Table I. RESULTS Complete Ascertainment A total of 262 individuals were evaluated for the presence of a psychotic illness based on their medical records or referral by a family member. Using RDC criteria, we identified 160 cases of strictly defined schizophrenia and 55 cases of other psychotic illnesses (see Table II). Included in the broad diagnostic category were five cases of acute/subacute schizophrenia, 31 cases of affective psychoses (22 bipolar affective disorder with psychosis, six schizoaffective disorder with mainly affective course, three psychotic depression), and 19 cases of unspecified functional psychosis. Of the remaining 47 individuals, eight were diagnosed as bipolar I without psychosis, bipolar II, or cyclothymic, 1 patient was diagnosed with delusional disorder, five TABLE II. Number of Subjects in Each Diagnostic Category by Gender Strict diagnostic category Chronic schizophrenia Schizoaffective, mainly schiz. Total strict schizophrenia Other psychotic illnesses Acute/subacute schizophrenia Affective Psychoses Bipolar I with psychosis Schizoaffective, mainly aff. Psychotic depression Unspecified functional psychosis Total other psychotic illnesses Broad diagnostic category Total all cases Male Female Total % Male 104 5 109 36 15 51 140 20 160 74.3 25.0 68.1 2 3 5 40.0 12 1 1 10 26 10 5 2 9 29 22 6 3 19 55 54.5 16.7 33.3 52.6 47.3 135 80 215 62.8 patients met criteria for psychosis but had a prominent organic component (e.g., chronic seizures), and 33 individuals (15 of whom were deceased) were considered unknown because of insufficient information. Among Palauans with strictly defined schizophrenia, there was a 2:1 male-to-female ratio with males representing 68.1% of total cases (chi-square ⳱ 12.54, df ⳱ 1, P < .01). This gender imbalance did not extend to patients with other psychotic illnesses, only 47.3% of whom are male. Prevalence We calculated the lifetime morbid risk (LMR) of schizophrenia based on complete ascertainment of cases and 1997 Palau population estimates by age group (Office of Planning and Statistics, Republic of Palau). The LMR [Gottesman and Shields, 1982] represents the probability that an individual who survives through the period of risk (15 to 54 years of age) will develop schizophrenia and is corrected for age using the following approximation to life table analysis: LMR = 冋 冉 a 1 N − n0 − nw 2 冊册 a = number of observed schizophrenics N = total sample size n0 = number of subjects under 15 years TABLE I. Diagnostic Categories Based on RDC Criteria Strictly defined schizophrenia Chronic schizophrenia (symptoms > 2 years duration) Subchronic schizophrenia (symptoms > 1 year but < 2 years duration) Schizoaffective disorder, mainly schizophrenic course Other psychotic illnesses Acute schizophrenia (symptoms < 6 months duration) Subacute schizophrenia (symptoms > 6 months but < 1 year duration) Schizoaffective disorder, mainly affective course Bipolar I disorder with psychotic symptoms Psychotic depression Unspecified functional psychosis nw = number of nonschizophrenics 15 to 54 years of age For strictly defined schizophrenia, the LMR in the Palauan population was 1.99%. When a broad diagnostic category including other psychotic illnesses was used, LMR for Palauans increased to 2.67%. For males, the risk of developing strictly defined schizophrenia was 2.77% and the risk of developing a broadly defined psychotic illness was 3.42%. The corresponding LMR estimates for females were substantially lower: 1.24% for strict schizophrenia and 1.95% for an illness in the broad diagnostic category. Schizophrenia in Palau, Micronesia 7 Age of Onset Table III presents mean age of onset by gender for each diagnostic category. The onset of strict schizophrenia typically occurred at an earlier age (24.7 ± 6.8 years) than the onset of other psychotic illnesses (28.8 ± 8.9 years), t ⳱ 3.02, df ⳱ 211, P < .01, two-tailed. Onset of strict schizophrenia in Palau occurred predominantly between ages 15 and 42 years (mean age of onset ⳱ 24.7 ± 6.8) with three males showing an earlier onset and one female showing a later onset. The average age of onset was 4 years younger for male (23.3 ± 6.2 years) than for female strict schizophrenia patients (27.5 ± 7.1 years), t ⳱ 3.79, df ⳱ 156, P < .001, twotailed. Age of onset distributions also showed strikingly different patterns by gender (see Fig. 1). There was a clear peak for Palauan men at 21–23 years of age. However, onset age for Palauan women showed a more complex pattern first peaking at 18 to 20 years, dropping back markedly at 20–23 years, then rising to a second peak that extends from 24 to 32 years of age. It is noteworthy that 61% of men with strict schizophrenia had onset before 24 years of age vs. only 29% of women, whereas 39% of women had onset in their thirties or later compared to only 18% of men. There was no gender difference in age of onset for patients with other psychotic illnesses. Familiality The 160 cases of strict schizophrenia represent a total of 59 families identified by a common founder. Eleven of these families each have 5 to 14 cases and represent nearly half (48.1%) of the strict schizophrenia patients in Palau, 20 families have 2 to 4 cases, 17 families have one patient who is distantly related to another previously counted case through a different common founder, and the remaining 11 strict schizophrenia patients are nonfamilial cases. Therefore, 93.1% of strict schizophrenia cases in Palau are familial (see Table IV). Of particular interest in this genetic isolate is the distribution of the illness within families. Although the 11 largest families each have 5 to 14 cases, these families cannot be described as high-density. Rather, the distribution of schizophrenia cases even within these multiply affected families would be more accurately described as sparse. To illustrate this point, Figure 2 presents a truncated pedigree drawing for Kindred 1583, which has the greatest number of cases descending from a single founder. The female founder had two children by two different males. The son had four children and the daughter had nine children, a total of 13 offspring in Generation III. These 13 individuals had a total of 94 offspring who comprise Generation IV. Among these 94 Generation IV family TABLE III. Mean Age of Onset in Years by Gender and Diagnostic Category Male Female Total Strict schizophrenia Other psychotic illnesses All cases combined 23.3 ± 6.2 27.5 ± 7.1 24.7 ± 6.8 28.7 ± 10.2 28.9 ± 7.6 28.8 ± 8.9 24.4 ± 7.4 28.0 ± 7.2 25.8 ± 7.6 Fig. 1. Frequency distribution of age of onset by gender. members, the youngest of whom are currently in their early 30s, there are only four affected cases to date. In Generation V, which currently ranges in age from 2 to 55 years and older, 10 cases of strict schizophrenia (one of which is counted in another family via the affected father) and three cases of Bipolar I with psychosis have been ascertained to date. In Generation VI, which will eventually span an even broader age range than Generation V, there is only one schizophrenia case so far. Another approach to describing the sparse distribution of cases within families is to examine number of cases by sibship size. Table V presents number of sibships with one, two, or three cases of strict schizophrenia by sibship size, including only full sibs. The average size of affected Palauan sibships is 5.7 individuals, and in these sibships, the risk for schizophrenia is 19.4%. Among the 160 cases ascertained, there are only 11 sib-pairs and two sib-trios. There are four additional half-sib-pairs in the sample. An important feature of the Palauan families segregating schizophrenia is the complex interconnections among the families via marriage that occur at every generational level studied. Based on these interconnections, the 48 multiply affected families we identified can be linked together into extended pedigrees with as many as 25 cases of schizophrenia. Figure 3, which presents a family identified by a single female founder TABLE IV. Cases of Strict Schizophrenia per Family 5–14 Strict cases per family 2–4 Strict cases per family Single-case families Familiala Nonfamilial Total a No. of families No. of cases Percent of total cases 11 20 77 55 48.1 34.4 17 11 59 17 11 160 10.6 6.9 100.0 Families with one strict schizophrenia patient who is related (third degree or more distant) to a previously counted case through a different common founder. 8 Myles-Worsley et al. Fig. 2. Pedigree for Kindred 1583 showing number of affected cases to date for each Generation III branch. with seven affected direct descendants, illustrates how marriages between clans can result in complex patterns of multilineal transmission. Of the eight marryin spouses that are circled, three are members of Kindred 1583 (presented in Fig. 2) and the other five are members of five other multiply affected families. Consequently, all seven affected descendants are either bilineal or multilineal. DISCUSSION The objectives of our epidemiological study were to estimate the prevalence of schizophrenia in Palau and TABLE V. Number of Sibships With 1, 2, or 3 Cases of Strict Schizophrenia Total no. of full sibs in sibship 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 Unknown Total no. of cases No. of sibships with 1 Affected 2 Affecteds 3 Affecteds 24 13 7 8 13 11 11 7 5 5 8 5 — 2 1 12 — — 1 1 — — 2 1 2 2 1 — — 1 — — — — — — — — 1 — — — 1 — — — — — 132 22 6 evaluate the influence of familial factors on the illness. Complete ascertainment of cases was possible because the relatively small population of Palauan adults lives in close geographical proximity in a communal society organized on the basis of large extended families or clans with excellent geneological documentation. The identification of potential cases was simplified by the centralization of all psychiatric care and medical records at Belau National Hospital and the assistance of the Palauan Behavioral Health staff who have extensive knowledge of patients and their families. We identified a total of 160 cases of strictly defined schizophrenia and estimated the lifetime morbid risk in Palau at 1.99%. Considering that Palau is a genetic isolate, this prevalence rate fits well within the range of estimates that have been reported in the literature over the past half century [Torrey, 1987; Jablensky et al., 1992]. It has been hypothesized that elevated prevalence rates in genetic isolates are due to the enrichment of genes predisposing to schizophrenia. However, enrichment would likely result in high-density pedigree branches, and these are rare in Palau. In Palau, expression of cases of schizophrenia resulting from susceptibility genes of low penetrance is more likely because of the large sibships and half-sibships produced by Palauan marriages. Another possible explanation for the elevated prevalence rate stems from the numerous marriages between members of different clans segregating schizophrenia. As a result of these interclan marriages at every generational level, a number of minor susceptibility genes may be aggregating in certain pedigree branches via a series of bilineal transmissions across multiple generations. Although the lifetime prevalence rate for schizophre- Schizophrenia in Palau, Micronesia 9 Fig. 3. Truncated pedigree for Kindred 2239B showing eight marry-in spouses (circled) and their family affiliation. nia may be somewhat elevated in Palau, this rate is not as high as prevalence rates found in other genetic isolates. For example, Hovatta et al.  recently reported an age-adjusted prevalence rate for schizophrenia in a Finnish isolate of 3.2%. The higher rate found in the Finnish isolate may be due to the broader diagnostic definition for schizophrenia which included schizophreniform disorder, all schizoaffective disorders, plus ‘‘simple’’ and ‘‘latent’’ schizophrenia. Also, cases were identified based on hospital registers which may have overdiagnosed schizophrenia: A test of their diagnostic validity showed that 13% of patients with a diagnosis of schizophrenia did not fulfill DSM-III-R criteria for the illness. It is also possible that we have understated the number of cases of schizophrenia in Palau. Our ‘‘unknown’’ category, which includes 36 individuals with some signs of psychiatric illness, likely includes some cases with premorbid symptomatology that could develop into schizophrenia over time as well as cases with atypical course that do not qualify as schizophrenia according to RDC criteria. A final possibility is that there are actual differences in the degree of elevation in prevalence rates across genetic isolates that result from the unique genetic characteristics of each isolated population. In Palau, the male-to-female risk ratio for schizophrenia is greater than 2:1. A similar finding was reported in the Finnish isolate where the age-adjusted lifetime prevalence rate was 3.9% for males and 2.4% for females [Hovatta et al., 1997]. These results are consistent with the majority of more recent epidemiological studies reporting an excess of male cases, including several studies in Ireland [NiNullain et al., 1987; Kendler et al., 1993], England [Cooper et al., 1987], Denmark [Munk-Jorgensen, 1986], Canada [Bland, 1984; Iacono and Beiser, 1992], and Korea [Lee et al., 1990]. We found that the mean age of onset for Palauans was 4.2 years earlier in men than in women. Kraepelin  first reported that men with a diagnosis of dementia praecox had been hospitalized an average of 5–10 years earlier than women. This observation of a gender difference in age at first admission has since been replicated in more than 50 studies [see An- germeyer and Kuhn, 1988 for a review]. However, no gender effects on age of onset were found in the Finnish isolate [Hovatta et al., 1997] nor in multiply affected families in County Roscommon, Ireland [Kendler and Walsh, 1995]. Furthermore, a recent reanalysis of data from the WHO 10-Country study of schizophrenia suggests that this gender difference in age of onset may be due, in large part, to differences in marital status [Jablensky and Cole, 1997]. It is important to note that these gender differences in age of onset apply to patients with strictly defined schizophrenia. For patients with other psychotic illnesses, which include affective psychoses, age of onset is no earlier in males than in females. The impact of family history factors on age of onset could not be evaluated because of the small number of nonfamilial cases in Palau. Schizophrenia in Palau clearly aggregates in families. However, the delineation of distinct families transmitting a heightened risk for schizophrenia is complicated by the multiple interconnections between pedigrees via marriage. We were able to assign the 160 cases of strictly defined schizophrenia to 59 families identified on the basis of a common founder. However, as interconnections between families were defined, it became apparent that the grouping of cases within families was often determined by which parent was ascertained first, particularly if that parent descended from one of the largest families. We found, for example, that it was possible to group eight cases from three different families under a different common founder. It has been difficult to find Palauan families that are exempt from psychotic illness. Most Palauans have at least a fourth degree relative who is affected. The well-documented genealogy of Palauan families that made complete ascertainment possible may account for the low proportion of nonfamilial cases (less than 7%) that we found in this isolated population. Family studies in large, outbred populations have reported a substantially higher proportion of sporadic cases [Gottesman, 1991]. However, if these families were as large and the genealogies as well known as they are in Palau, many of these ‘‘sporadic’’ cases might actually be redefined as distantly familial. 10 Myles-Worsley et al. With complete ascertainment of schizophrenia in Palau accomplished, our study is now proceeding with the search for susceptibility genes in this geographic and ethnic isolate. As in the Finnish isolate, the majority of the ascertained cases can be linked together into extended pedigrees with complex multilineal patterns of transmission of the clinical phenotype. Although traditional analytical techniques may not be capable of generating strong linkage signals in the more complex multilineal pedigrees, these Palauan families represent a potentially valuable data resource for the application of new analytical techniques based on neural network models [Kaufman, 1997] that can accommodate complex mechanisms of inheritance such as gene interactions and epistasis. ACKNOWLEDGMENTS This research was supported by grants MH54186 (M.M.W.), MH56098(W.B.), MH10889(W.B.), and MH52055 (H.C.) from the National Institute of Mental Health, a NARSAD Independent Investigator Award (M.M.W.), and a NARSAD Senior Investigator Award (W.B.). We greatly appreciate the support and cooperation of the families in Palau. We thank Leatha Allred, Patrick Aribuk, Adeline Franz, Annabel Lyman, Humi Masayos, Felix Mereb, Hilda Ngiralmau, Druscilla Patris, Jennifer Williams, and the support staff at Belau National Hospital for their important contributions. REFERENCES Angermeyer MC, Kuhn L. 1988. Gender differences in age at onset of schizophrenia: An overview. Eur Arch Psychiatr Neurol Sci 237:351– 364. Bland RC. 1984. Long-term mental illness in Canada: An epidemiological perspective on schizophrenia and affective disorders. Can J Psychiatry 29:242–246. Book JA, Wetterberg L, Modrzewska K. 1978. Schizophrenia in a north Swedish geographical isolate, 1900–1977. Epidemiology, genetics and biochemistry. Clin Genet 14:373–394. Cooper JE, Goodhead D, Craig T, Harris M, Howat J, Korer J. 1987. The incidence of schizophrenia in Nottingham. Br J Psychiatry 151:619– 626. Dale PW. 1981. Prevalence of schizophrenia in the Pacific Island populations of Micronesia. J Psychiatr Res 16:103–111. Endicott J, Spitzer RL. 1978. A diagnostic interview: The schedule for affective disorders and schizophrenia. Arch Gen Psychiatry 35:837– 844. Gottesman II. 1991. Schizophrenia Genesis. New York: WH Freeman. Gottesman II, Shields J. 1982. Schizophrenia: The epigenetic puzzle. New York: Cambridge University Press. Gottesman II, McGuffin P, Farmer AE. 1987. Clinical genetics as clues to the ‘‘real’’ genetics of schizophrenia: (A decade of modest gains while playing for time). Schizophr Bull 13:23–47. Hovatta L, Terwilliger J, Lichtermann D, Makikyro T, Suvisaari J, Peltonen L, Lonnqvist J. 1997. Schizophrenia in the genetic isolate of Finland. Am J Med Genet 74:353–360. Iacono WG, Beiser M. 1992. Are males more likely than females to develop schizophrenia? Am J Psychiatry 149:1070–1074. Jablensky A, Sarorius N, Ernberg G, Anker M, Korten A, Cooper JE Day R, Berrtelsen A. 1992. Schizophrenia: Manifestations, incidence and course in different cultures. Psychol Med 22[Monograph Suppl 20]. Jablensky A, Cole S. 1997. Is the earlier age at onset of schizophrenia in males a confounded finding? Results from a cross-cultural investigation. Br J Psychiatry 170:234–240. Kaufman CA. 1997. Neural network analysis of schizophrenia susceptibility loci. Presented at the World Congress on Psychiatric Genetics, Santa Fe, NM, October 1997. Kendler KS, McGuire M, Gruenberg AM, O’Hare A, Spellman M, Walsh D. 1993. The Roscommon family study. I. Methods, diagnosis of probands and risk of schizophrenia in relatives. Arch Gen Psychiatry 50:527– 540. Kendler KS, Walsh D. 1995. Gender and schizophrenia: Results of an epidemiologically-based family study. Br J Psychiatry 167:184–192. Kraepelin E. 1909. Psychiatrie. Leipzig: Barth. Lee CK, Kwak YS, Yamamoto J, Rhee H, Kim YS, Han JH, Choi JO, Lee YH. 1990. Psychiatric epidemiology in Korea. Part I: Gender and age differences in Seoul. J Nerv Ment Dis 178:242–246. Munk-Jorgensen P. 1986. Schizophrenia in Denmark: Incidence and utilization of psychiatric institutions. Acta Psychiatr Scand 73:172–180. NiNullain M, O’Hare A, Walsh D. 1987. Incidence of schizophrenia in Ireland. Psychol Med 17:943–948. Parmentier RJ. 1987. The sacred remains: Myth, history, and polity in Belau. Chicago: University of Chicago Press. Simmons RT, Graydon JJ, Gajdusek DC, Brown P, Riesenberg SH. 1965. Blood group genetic variations in natives of the Caroline Islands and in other parts of Micronesia. Oceania 36:132–170. Spitzer RL, Endicott J, Robins E. 1978. Research diagnostic criteria. Arch Gen Psychiatry 35:773–782. Takayama J. 1981. Early pottery and population movements in Micronesian prehistory. Asian Perspectives 24:1–10. Torrey EF. 1987. Prevalence studies in schizophrenia. Br J Psychiatry 150:598–608.