Exfoliation syndrome: prevalence and inheritance in a subisolate of the Finnish population

Authors


Eva Forsman
Västerbackavägen 4
10120 Täkter
Finland
Tel: + 358 40 5757246
Fax: + 358 90 2214413
Email: eforsman@elisanet.fi

Abstract.

Purpose:  To estimate the prevalence of pseudoexfoliation syndrome or exfoliation syndrome (ES) in a cross-sectional study and during a long-term follow-up, and to analyse how ES has been inherited in a large pedigree on an isolated population of Kökar island in southern Finland.

Methods:  In a population-based study conducted between 1960 and 1962, a comprehensive ophthalmological examination was performed on 595 subjects (85% of the population). From then until 2002, 965 subjects were examined at least once. A pedigree was constructed for all ES-affected subjects according to genealogical studies. The genetic contribution to ES was investigated in this pedigree by segregation analysis and the heritability of the intraocular pressure (IOP) quantitative trait estimated using SOLAR and SAGE software.

Results:  In the cross-sectional study, the prevalence of ES was 8.1% for 247 subjects over 50 years of age (males 7%, females 9%) and increased to 18.4% for 70 subjects over 70 years of age (males 13%, females 25%). In addition, two females less than 50 years of age were ES-affected. Between 1960 and 2002, 76 (14.3%) of 530 subjects over 50 years of age had ES [23 males (10%) and 53 females (18%)]. Exfoliation glaucoma (EG) was found more often in males (11 patients, 48%) than in females (13 patients, 25%) whereas primary open-angle glaucoma (POAG) was almost as frequent in males (seven patients, 3%) as in females (five patients, 2%). The relative risk (RR) of glaucoma (ES versus no ES) was 11.9 [95% confidence interval (CI) 6.2–22.9] for all the subjects – 14.6 for males (95% CI 6.3–34.0) and 11.8 for females (95% CI 4.4–31.6). Seventy-five pedigrees of 78 ES-affected patients were linked together into one large pedigree with 110 nuclear families. The segregation ratio of ES was 18% (8% for males, 24% for females) when both parents were unaffected, and 16% (9% for males, 27% for females) when at least one parent was affected. The heritability of IOP was estimated to be 30%.

Conclusion:  In this population-based family study, ES is consistent with an autosomal dominant trait with incomplete penetrance, where the penetrance is more reduced in males than in females. However, the presence of ES was a greater risk factor for developing glaucoma in males than in females.

Introduction

Pseudoexfoliation syndrome or exfoliation syndrome (ES) was first described by John Lindberg (1917) as ‘whitish-grey scales which emanated from the outermost part of the pupillary border and occasionally grouped in a more or less dense ring on the anterior capsule of the lens’. Since then, ES has been confirmed as a risk factor for developing open-angle glaucoma (Klemetti 1988; Ekström 1993; Mitchell et al. 1999), for glaucoma progression (Aasved 1971; Pohjanpelto 1986; Leske et al. 2003) and for more severe prognosis in glaucoma (Aasved 1971; Pohjanpelto 1986; Brooks & Gillies 1988; Konstas et al. 1998; Vesti & Kivelä 2000). The prevalence of ES varies widely among different ethnic populations from zero to around 20–25% (Forsius 1988; Ringvold et al. 1991; Hirvelä et al. 1994; Foster et al. 1996; Kozobolis et al. 1997; Mitchell et al. 1999; Ringvold 1999). It is associated strongly with increasing age; prevalence is highest between the ages of 70 and 80 (Krause et al. 1988; Ringvold et al. 1991; Kozobolis et al. 1997; Mitchell et al. 1999). Because ES has often been found clustered in families, genetic factors have been proposed to play an important role in its aetiology (Allingham et al. 2001). In a Finnish population-based twin study with chronic open-angle glaucoma and exfoliation glaucoma (EG), the heritability of EG could not be assessed by the classical twin method because of the small number of pairs (Teikari 1987). In an Icelandic twin study, five of eight monozygotic twins with exfoliation were concordant for ES with their cotwin, supporting the role of genetics in the aetiology underlying ES (Gottfredsdottir et al. 1999). A few studies (Tarkkanen 1962; Pohjanpelto & Hurskainen 1972; Aasved 1975; Damji et al. 1999; Allingham et al. 2001; Orr et al. 2001) in families consisting of between two and four generations have been described. Depending upon the study material, different modes of inheritance have been suggested: autosomal dominant (Tarkkanen 1962; Aasved 1975; Orr et al. 2001), autosomal recessive (Andersen et al. 1997) and maternal transmissions such as X-linked, mitochondrial and autosomal inheritance with genomic imprinting (Damji et al. 1999). The main problem for investigations of the genetic background of ES has been the late onset of ES, where previous generations are often deceased at the time when the current generation is manifesting exfoliative changes.

In earlier studies, an affected ES-positive proband has been identified in an eye clinic or a private practice followed by an invitation to family members to participate in a clinical study. In the present study, the affected subjects were drawn from a population-based study conducted between 1960 and 1962 and a follow-up of the same population during the next 40 years. The hypothesis was that ES could have been enriched in this population because of the isolation of the population during centuries on Kökar, an outer parish in Åland archipelago on the south-western coast of Finland. The aims of this investigation were to evaluate the prevalence of ES in a cross-sectional study and during a long-term follow-up, and to model how ES has been transmitted.

Materials and Methods

A population-based study was carried out by two of the authors (H.F. and A.E.) between 1960 and 1962 on the isolated island of Kökar, which has been permanently populated since the 12th century (Eriksson & Forsius 1964). The primary aim of the study was to find out the prevalence of retinitis albicans punctata and von Willebrand's disease, which were both prevalent in Åland (Eriksson et al. 1964; Forsius & Eriksson 1964). Altogether, 595 inhabitants over 10 years of age – 85% of the whole population – were examined ophthalmologically (Fig. 1). During the subsequent decades, follow-up studies of the same population, their relatives and those who had not participated the first examination were organized. Studies were conducted six times between 1975 and 1995 (50–200 participants), in 1991–92 (about 400 participants) and most recently in 2001–02 (183 participants). In total, 965 subjects, 530 of them older than 50 years, were examined at least once.

Figure 1.

 The study population in Kökar. ES, exfoliation syndrome; +, affected; –, unaffected; EG, exfoliation glaucoma; POAG, primary open-angle glaucoma.

Clinical data included medical and ocular history, gender, visual acuity, intraocular pressure (IOP) [except when glaucoma was diagnosed and treated prior to the study examination(s) in seven individuals], slit-lamp biomicroscopy and funduscopy before and following dilation.

ES was defined by the presence of a greyish central disc with or without focal breaks with or without a peripheral band on the anterior lens capsule and/or fibrillary material on the pupillary ruff, and recorded without grading. A subject was recorded as an affected ES-positive if ES was detected in at least one eye. Suspect changes such as Krukenberg's spindle and pigmentation of the cornea endothelium, or a diffuse haze on the anterior capsule, were noted as negative. The age of subjects included in prevalence estimates and in the pedigree analysis was either the age when ES was observed for the first time or (if ES was not present) the age when the last examination was performed. Changes in the optic nerve head as cup-to-disc ratios (C/D) of 0.5–0.6 or more, and/or asymmetry in the C/D between the right and left eyes and/or elevated IOP > 21 mmHg arouse a suspicion of glaucoma. All those patients were referred to the Central Hospital of Åland, where the diagnosis of primary open-angle glaucoma (POAG) or EG was confirmed based on clinical evaluation and visual fields. A large part of the study population was examined at the time when gonioscopy was not available; therefore, the width or pigmentation of the anterior chamber angle was not evaluated.

From the inception of the study, the names, birth dates and places of birth of parents, grandparents and great grandparents were collected from all participants (Eriksson & Forsius 1964), using the same questionnaire throughout the whole study. The names and dates and places of the birth were utilized to trace participants' ancestors from local church registers – the only one on the island because membership of the (Lutheran) Church was obligatory until 1920 in Finland. Population records kept by the Finnish Church since the 17th century provide precise details of births, deaths and marriages (Varilo et al. 1996). The collected information was combined with church registers, resulting in 110 nuclear families who came from one large pedigree.

The study was approved by the Ethical Committee of the Central Hospital of Åland in the early 1990s and in 2001. All participants gave their informed consent.

Segregation analysis of ES and heritability estimate of IOP

The genetic contribution to ES was investigated by a segregation analysis conducted on the nuclear families contained within the larger pedigree, where the parents had a known ES status (Jarvik 1998). The configurations of ES-affected and -unaffected male and female offspring are stratified by the ES affected, unaffected and unknown status of their parents in Table 1. To estimate the segregation ratios in samples of a sufficient size, the parental mating types were stratified into matings of (1) unaffected by unaffected and (2) at least one parent affected, with the second parent having any status. Within these two categories, the segregation ratios of the male and female children were estimated, combined and separately, by dividing the number of affected offspring by the total number of offspring in each category.

Table 1.   Segregation table.
Mating type
(father × mother)
Nuclear families (n)Children (n)Affected children (n)
FemaleMaleFemaleMale
  1. U, unknown; A, affected by exfoliation syndrome; N, not affected by exfoliation syndrome.

U × U22221663
N × U84530
A × U41700
U × N18162694
N × N284226102
A × N43811
U × A17171552
N × A9101231
A × A32400
Total1101171193713

An analysis to estimate the heritability of the IOP quantitative trait in the entire pedigree of 309 individuals with trait values was conducted using the SOLAR software (Almasy & Blangero 1998). This variance components analysis of a normally distributed quantitative trait models the trait covariances among the relatives in the pedigrees, according to their degrees of relationships, and uses this information to derive the best estimate of heritability. This estimate is confounded by the effects of a common family environment, which can inflate it, making the heritability estimate an upper bound. To further interpret the result of this analysis, correlations among the different classes of relative pairs in the nuclear families within the pedigree were estimated separately using the familial correlations (FCOR) subprogram of the Statistical Analysis for Genetic Epidemiology (SAGE) software package. This was carried out for male–male, male–female and female–female sibling pairs to better detect any effect of sex on the heritability of IOP.

Results

Population-based study, 1960–62

Altogether 595 inhabitants over 10 years of age participated in the study (Fig. 1). Two females under the age of 50 had ES (one with EG). Of all the subjects examined, 247 were older than 50 (89 males, 158 females), mean age 64. The prevalence of ES among them was 8.1%. Twenty were affected by ES; of these, six were male (frequency 7%) and 14 were female (frequency 9%). The prevalence increased with age and was 18.4% among 76 individuals aged ≥ 70 years (males 13%, females 25%). Three males had EG and one male had POAG. The odds ratio (OR) of glaucoma (ES versus no ES) was 40 [95% confidence interval (CI) 4–404] for all and 82 (95% CI 7–1039) for males.

The population, 1960–2002

At the end of 2002, a total of 965 subjects over 10 years of age had been examined ophthalmologically. At their last examination, 530 subjects (237 males, 293 females) were over 50 years of age (mean age 63.3). Altogether, 76 (14.3%) subjects with ES [23 males (10%), 53 females (18%)] were identified from the beginning of the study (Table 2). Unilateral and bilateral exfoliations were present in equal proportions. The presence of ES was associated with age and was most prevalent between 70 and 79 years in both genders. The affected individuals were older than the unaffected participants. The mean ages of males with and without ES were 72 years and 65 years, respectively. The mean ages of females with and without ES were 72 and 66 years, respectively. The mean IOP in subjects with ES without glaucoma was 17 mmHg (range 10–27), with glaucoma (EG) 28 mmHg (range 16–52). The mean IOP in unaffected subjects without glaucoma was 17 mmHg (8–29), with glaucoma (POAG) 26 mmHg (range 16–38). In unilateral cases, the affected eye had a mean IOP of 19 mmHg (range 10–47) and the unaffected eye 16 mmHg (range 11–24). EG was found more often in males (11/48%) than in females (13/25%), whereas POAG was almost as frequent in males (7/3%) as in females (5/2%). The relative risk (RR) of glaucoma (ES versus no ES) was 11.9 (95% CI 6.2–22.9) for all, 14.6 (95% CI 6.3–34.0) for males and 11.8 (95% CI 4.4–31.6) for females.

Table 2.   Exfoliation syndrome (ES) and exfoliation glaucoma (EG) in the population.
 Age 50–59 yearsAge 60–69 yearsAge 70–79 yearsAge ≥ 80 yearsMean
total
(n = 530)
Male
(n = 68)
Female
(n = 90)
Male
(n = 83)
Female
(n = 85)
Male
(n = 60)
Female
(n = 75)
Male
(n = 26)
Female
(n = 43)
  1. ES–, unaffected by ES; ES+, at least one eye affected by ES; unilateral, ES in either eye; bilateral, ES in both eyes.

ES–68 (100%)86 (96%)72 (87%)70 (82%)50 (83%)52 (69%)24 (92%)32 (74%)454 (86%)
ES+04 (4%)11 (13%)15 (18%)10 (17%)23 (31%)2 (8%)11 (26%)76 (14%)
Unilateral ES+02585120638
Bilateral ES+02675112538
Individuals with EG0252551424

Family studies

In accordance with published criteria (Varilo et al. 1996), genealogical studies were performed to construct a pedigree for each of the 78 affected subjects including the two females in whom ES was detected before 50 years of age. Three ES-positives (two males, one female) of all affected individuals had moved from neighbouring parishes and thus were not genuine inhabitants of Kökar. Interestingly, all the small pedigrees were linked together into one large pedigree of 13 original families in six to eight generations (Fig. 2). The pedigree included 332 examined subjects over 50 years of age (160 males, 173 females). The remaining 188 of 530 examined subjects were far more distant relatives and were not included in the pedigree. Within the large pedigree, ES was present in 22% of the examined relatives compared to 14.5% in the whole population, suggesting a genetic predisposition. However, the sibling recurrence risk (Ks) for siblings over 70 years of age in this extended family was 15% and the sibling recurrence risk ratio (λs) 0.8 [calculated according to the formula suggested by Farbrother et al. (2004)]. The mean age of the unaffected individuals in the pedigree was slightly higher than that of unaffected individuals who were not included in the pedigree (67 years old versus 63 years old). Glaucoma was detected in 31 family members; of those, 23 had EG (10 males, 13 females) and eight had POAG (seven males, one female).

Figure 2.

 Pedigree of 75 individuals with exfoliation syndrome (ES) in Kökar. The affected individuals are represented by black symbols and the generations by Roman numerals.

Nearly 90% of family members over 50 years of age in the three last generations participated in the study over the course of 40 years. Twenty-seven per cent of the sixth generation (27 of 100 examined subjects) and 25% of the seventh generation (37 of 146 examined subjects) were affected – even though the younger generation had been examined more often than the previous generation. The mean age of the last (eighth) generation was only 59 years, with 11 affected.

Transmission

The pedigree included 110 nuclear families with parent–offspring pairs. Each mother and father was split into subgroups: unaffected ES-negatives, affected ES-positives and unexamined (Table 3). About one third of the parents in each subgroup had at least one affected offspring, except the ES-positive fathers subgroup, in which only one had affected offspring (a daughter aged 59 years and a son aged 64). However, in this case the unaffected mother's sister was also affected (Fig. 3). The 10 other fathers had five daughters without ES (one aged 80 years and four aged 49–58 years) and 18 sons without ES (aged 52–94 years). Both parents were affected in three of the cases; surprisingly, all six offspring were ES-negative, but five of them were relatively young, with ages between 49 and 61 at examination (Table 1). The proportion of ES-positives among males (11%) was much lower than among females (32%), although the total number of investigated male and female offspring was nearly the same. A clustering of ES within the large pedigree was also evident in some families. One of these included 26 affected individuals; in this family, the mode of inheritance was consistent with autosomal dominant in three generations (Fig. 4).

Table 3.   Exfoliation syndrome (ES) in offspring.
 Female (n = 117)Male (n = 119)Total (n = 236)Siblings (n = 110)
ES–
[n = 80 (68%)]
ES+
[n = 37 (32%)]
ES–
[n = 106 (89%)]
ES+
[n = 13 (11%)]
ES–
[n = 186 (79%)]
ES+
[n = 50 (21%)]
ES–
[n = 74 (62%)]
ES+*
[n = 36 (38%)]
  • *

    At least one of the siblings was affected.

  • ES–, unaffected individual; ES+, affected individual.

Mother (n = 110)
 ES–[n = 50 (45%)]41 (67%)20 (38%)52 (88%)7 (12%)94 (78%)27 (22%)34 (68%)16 (32%)
 ES+ [n = 26 (24%)]21 (72%)8 (28%)28 (90%)3 (10%)49 (82%)11 (18%)17 (65%)9 (35%)
 Unknown [n = 34 (31%)]18 (68%)9 (33%)25 (89%)3 (11%)43 (78%)12 (22%)23 (68%)11 (32%)
Father (n = 110)
 ES–[n = 45 (41%)]40 (71%)16 (29%)40 (93%)3 (7%)80 (81%)19 (19%)29 (64%)16 (36%)
 ES+ [n = 11 (10%)]5 (83%)1 (17%)18 (95%)1 (5%)23 (92%)2 (8%)10 (91%)1 (9%)
 Unknown [n = 54 (49%)]35 (64%)20 (36%)48 (84%)9 (16%)83 (74%)29 (26%)35 (65%)19 (35%)
Figure 3.

 Pedigree of the father with exfoliation syndrome (ES) (indicated by an arrow) and his affected offspring. ES is represented by the black upper half of the symbol. The unexamined individuals are represented by a question mark. The age of diagnosis (for affected individuals) or age at last examination (for unaffected individuals) is below the symbol. The generations are represented by Roman numerals.

Figure 4.

 Pedigree of the family with exfoliation syndrome (ES). Totally black symbols denote exfoliation glaucoma (EG), symbols with a black upper half denote ES, symbols with a black lower half denote primary open-angle glaucoma (POAG) and a question mark denotes an unexamined individual. The age of diagnosis (for affected individuals) or age at last examination (for unaffected individuals) is represented below the symbol. The generations are represented by Roman numerals.

Segregation ratio of ES and heritability of IOP

The segregation ratio of ES was estimated in the offspring from the nuclear families in the large pedigree. These nuclear families were stratified into two classes of parental mating types: (1) diagnoses of unaffected by unaffected in the parents and (2) at least one parent diagnosed with ES. These families are summarized in Table 1. In the 28 nuclear families with both parents unaffected, there were 12 affected offspring out of 68, resulting in a segregation ratio of 18%, and the male and female segregation ratios were 2/26 (8%) and 10/42 (24%), respectively, indicating a marked difference in segregation ratios between males and females, with a much higher ratio in females. In the 34 matings where at least one parent was diagnosed with ES, there were 13 affected children out of 79, resulting in a segregation ratio of 16%. When the offspring were stratified by sex, the male and female segregation ratios were 4/46 (9%) and 9/33 (27%), respectively, resulting in segregation ratio estimates that are very similar to those observed in the offspring of two unaffected parents.

Heritability is an estimate of the percentage of the variance of a trait that can be attributed to genetic effects. This can vary according to changes in the relative contributions of genes and environment to IOP for different classes of individuals, such as males and females. The heritability of IOP was estimated to be 30% when the entire pedigree was analysed intact using the SOLAR software and not taking the sex of the individual into consideration. However, we also calculated separate pairwise correlations of the 52 male–male, 119 male–female and 75 female–female sibling pairs. If heritability is the same across the sexes, these correlations should be consistent; however, they are 0.02, 0.13 and 0.56, respectively. These marked differences imply that there is likely to be an effect of the sex of the individual on the degree of the genetic contribution to this trait. From these analyses, we infer that genes have a much greater effect on the variance of IOP in females than in males.

Discussion

We have conducted a population-based longitudinal study of ES and IOP in a Finnish cohort from a single island in the Gulf of Bothnia. In the cross-sectional study of data collected in 1960–62, the prevalence of ES was 8.1% among 247 subjects over 50 years of age and 18.4% among 76 subjects over 70 years of age. During a 40 year follow-up of the same population, 530 individuals over 50 years of age were examined. ES was detected in 18% of females compared to 10% of males. The prevalence of ES has been quite similar in all Scandinavian population-based studies (Ekström 1987; Krause et al. 1988; Ringvold et al. 1991; Hirvelä et al. 1994). In the present study, the prevalence among those over 70 years of age (18.4%) was a little less than that reported in a population-based study from Oulu in Northern Finland (22%) (Hirvelä et al. 1994). However, there was no significant difference in prevalence figures among males (12.5% in Kökar versus 13% in Oulu) and among females (25% versus 27%, respectively). In a population-based study carried out in Mid-Norway, the corresponding figures were somewhat lower for the whole population (16.6%) and for females (18.2%) but not for males (14.6%) (Ringvold et al. 1991) when the inclusion age for the study was 64 years or over. When compared to these two population investigations, our results show that the ES trait has not been enriched in the isolated population of Kökar.

ES expresses itself in various ways. The amount of exfoliation material varies enormously and secondary changes (such as glaucoma) develop for approximately 40% of patients (Klemetti 1988). In the present study, the proportion of EG in males (48%) was considerably higher than that in females (25%). The same variation was observed in RR values: 14.6 (95% CI 6.3–34.0) and 11.8 (95% CI 4.4–31.6), respectively. Ringvold et al. (1991) reported a sex difference in proportions of subjects with EG in a Mid-Norway study (35.9% in males versus 25.7% in females), as did Ekström (1993) in a Swedish study (RR of 30 for males versus 10 for females). The same tendency was observed in six Icelandic families with 25 ES-affected members, of which seven males (63%) and two females (21%) had EG (Allingham et al. 2001). The study material has often been sex-adjusted, and therefore ES has not been graded to be a greater risk factor for glaucoma in males than in females.

Ocular hypertension without glaucomatous changes in the optic disc or visual fields is a common finding, but simultaneously elevated IOP is a major risk factor for developing open-angle glaucoma (OAG). Although most different glaucoma types have been shown to be genetic, only one study concerning the genetic contribution of IOP has been published. Duggal et al. (2005) conducted a heritability analysis similar to the one we report here of 2337 individuals from 620 pedigrees in the Beaver Dam Eye Study. Remarkably, their overall heritability estimate for IOP (30%) was the same as ours. They also estimated family correlations using the same FCOR software we used in our analysis, but did not stratify the sibling pairs by their sex. Their estimate of the IOP correlation for 1125 sibling pairs was 0.15, while ours was larger at 0.36 in 242 sibling pairs. However, when we stratified the pairs by sex we were able to see an intraclass correlation of 0.56 in the female pairs and a much lower value for male pairs, revealing that the gene or genes contributing to aberrant IOP values are more likely to be penetrant in females, leading to their higher sibling correlation. The higher sibling correlation in our sample could be the result of a larger number of female sibling pairs in our sample.

These heritability estimates and segregation ratios are consistent with a dominant model of inheritance of ES susceptibility with reduced and different penetrance values in males and females. Most of the fathers and some of the mothers in the matings of two unaffected parents are likely to be carriers of a gene or genes predisposing to ES. However, these parents do not develop the disorder because of reduced penetrance, but pass the genes on to their offspring in a fashion similar to that observed when at least one parent is affected. Thus, we observe very similar segregation ratios in both types of families where they pass a gene to 50% of their offspring, who also exhibit reduced penetrance that is dependent on sex. From these analyses, the penetrance of the gene may be as low as 10–15% in males but as high as 50% in females. This is consistent with the heritability estimates of the IOP quantitative trait, which is high in females and very low in males. The difficulties in these analyses arise from the late onset of ES, where everyone is a possible carrier but the gene remains non-penetrant. In most studies (Ekström 1987; Ringvold et al. 1991; Hirvelä et al. 1994), the prevalence of ES has been higher among females than among males – as in this study – except in some series published from Greece (Kozobolis et al. 1997) and the tropics (Forsius et al. 2002). Consistent with our findings, ES has previously been proposed to be inherited as an autosomal dominant trait with incomplete penetrance. In our study, support for this model was found in the segregation analysis of siblings from different mating types.

There has been only one reported study where fathers have been observed to be transmitters of the ES trait to their offspring (Orr et al. 2001). The mothers were transmitters in families studied by Damji et al. (1999) and Allingham et al. (2001); therefore, X-linked, mitochondrial or genome imprinting models have been proposed. With X-linked inheritance, a son cannot inherit the trait from his father. In dominant mode, all daughters of an affected father should be affected, and there are twice as many affected females as males (unless the disease is lethal in males), whereas in the recessive mode all daughters of an affected father are heterozygous healthy carriers and there is greater prevalence of affected males (Jorde et al. 1999). In our study, ES was prevalent in 18% of females and 10% of males > 50 years of age. However, among the offspring the female : male ratio was 3 : 1. Only one of the 11 affected fathers (Fig. 3) had an ES-positive male offspring, but in that case the mother's sister was also affected providing the possibility that the trait had been inherited from the mother – thus indicating that X-linked inheritance has not been ruled out for ES. Mitochondrial inheritance would have only maternal transmission, and is thus possible if the trait is inherited from the mother's side in this family. Imprinting would result in a different phenotype if the trait was inherited from the mother or father, which is again consistent with the ambiguous inheritance pattern in this family. In addition, one of the six daughters of the affected fathers was affected (Tables 1 and 3), and therefore dominant X-linked inheritance could be a possibility because the ages of four daughters were between 49 and 58 years. Based on the results shown in this article, there is no confirmed mode of inheritance in ES.

In summary, the variation of the phenotype and clustering in families support a model of oligogenic or even polygenic inheritance with a likely environmental component. Genetics were discussed in an earlier article (Forsius et al. 1993); however, in the larger, more formally analysed sample reported here, an autosomal dominant model with incomplete penetrance is the most likely feature of inheritance. Penetrance is more reduced in males than in females, but the presence of ES provides a greater risk factor for glaucoma in males than in females. Our study further emphasizes the role of genetic factors in the pathogenesis of ES and the need for molecular genetic studies.

Acknowledgements

This work was supported by grants from the Sigrid Juselius Foundation, Finland and the Eye and Tissue Foundation, Finland.

Ancillary