Dermatoglyphs of Digito-Palmar Complex in Autistic Disorder: Family Analysis

Transcription

1 44(4): ,2003 CLINICAL SCIENCES Dermatoglyphs of Digito-Palmar Complex in Autistic Disorder: Family Analysis Jasna Milièiæ, Zorana Bujas Petkoviæ 1, Jadranka Bo ikov 2 Institute for Anthropological Research, 1 Jankomir Psychiatric Hospital, and 2 Andrija Štampar School of Public Health, Zagreb, Croatia Aim. To examine the role of the genetic component in the quantitative of autistic patients and their families, and the transmission of the autism. Methods. Finger and palm prints were taken from 120 autistic patients (92 males and 28 females), their parents (92 mothers and 70 fathers), 32 healthy brothers and 28 sisters, as well as 400 healthy controls (200 males and 200 females). An analysis of quantitative of dermatoglyphs on the fingers (FRC finger ridge count) and palms (a-b, b-c, and c-d ridge count, and atd angle) was performed. Descriptive statistics, multivariate analysis of variance (MANOVA) with Tukey HSD post hoc test, and discriminant analysis were used to determine the differences among the groups. In addition, correlations among family members were analyzed. Results. Multivariate analysis showed significant differences among examined groups of autistic patients and their family members and healthy volunteers regarding both group membership and sex. Autistic male patients differed significantly from the healthy controls in the ridge count (RC) on the fourth and fifth finger, and in and atd angle of both hands. Healthy fathers of autistic patients differed in atd angle, and brothers of autistic patients differed in all palmar variables from the healthy control group. Mothers of autistic patients differed significantly from the healthy female controls in the RC of the first, fourth, and fifth finger, in a-b and on the palms, and atd angle of both hands. The first two discriminant functions explained 85.4% of variance and separated groups clearly in two ways: the first function separated healthy controls from family members of autistic patients, and the second one males from females. Interfamilial analysis showed significant interclass correlations between autistic sons and their mothers or fathers in practically all variables. However, the correlation between parents and their autistic daughters was lower. Both mothers and fathers of autistic patients correlated with their healthy children only in palmar variables. Conclusions. We found significant differences in ridge counts on the fingers and palms between the affected patients and their healthy controls, but these differences also existed between family members of autistic patients and healthy controls. Particularly pronounced were the differences between healthy female controls and female family members, including not only autistic female patients, but also their healthy mothers and sisters. Since the mothers and their autistic sons showed higher statistically significant correlation in most of the examined variables, unlike the mothers and their autistic or healthy daughters, it is possible that there is a connection between a recessive X-chromosome linkage, as a genetic component in the etiology of autistic disorders, and the influence of the inactivation of the affected X-chromosome in the females. Key words: autistic disorder; dermatoglyphs; fingers; genetic predisposition to disease; inheritance patterns; X-chromosome Autism is a severe neurodevelopmental disorder characterized by communication and social deficits and stereotyped, repetitive behavior (1). The syndrome of autism is highly heritable and considered to be etiologically heterogeneous. It is also thought to be the result of multiple, interacting genes (2,3). There is a general agreement that genetic factors play a major role in the predisposition to autism (4). Numerous studies have shown that the disorder occurs 50 to 100 times more frequently within some families than in the general population (5). The concordance rate for monozygotic twins is 50-80%, but only 3% for dizygotic twins and brothers and sisters of autistic patients (6). Autism occurs in combination with some genetic diseases, such as tuberous sclerosis (7), phenylketonuria (8), and neurofibromatosis (9). Although structural and functional temporal lobe abnormalities seem important in autism, in most cases they appear as the consequence of the abnormalities of neuronal organization, connectivity, and function of the brain (10). The most important findings are micropathologic abnormalities in the cerebellum, consistently found in association with limbic system abnormalities in autistic individuals. These abnormalities probably 469

2 develop before the 30th week of gestation, simultaneously with the development of finger and palm ridges (11,12). So far, no specific gene for autism has been identified, although a possible link with the serotonin-transporter gene has recently been suggested (13,14). An important observation is the association of the autistic disorder with a fragile X syndrome (15,16). Genetic screening of individuals with autism revealed the incidence of fragile X syndrome from 0% (1) to 12% or higher (15,17). During the early embryonic development, pathological genetic factors, which disrupt the development of the central nervous system, could also influence the early formation of dermatoglyphs in autistic patients (12). This is why the study of the dermatoglyphs may offer some insight into the events during fetal development (18). Dermatoglyphs are cutaneous ridges on the fingers, palms, and soles, formed during early intrauterine life, between the 7th and 21st week of gestation (19). During this period, and only then, genetic and environmental factors can influence their formation (20-22). Sometimes, genetic influence can drastically change the dermatoglyphs, causing no flexion creases on the palms of affected children (Fig. 1). Therefore, dermatoglyphs are used as easily accessible tool in the study of genetically influenced diseases. Several studies of dermatoglyphs of the digitopalmar complex in autistic patients have been published (Table 1). Walker (18) reported on the differences in the quantitative and qualitative characteristics of the dermatoglyphs between autistic children and healthy examinees, with observable diminishment of dermal ridge development with a deviancy of ridge patterns in autistic patients. Hartin and Barry (23) emphasized the differences among autistic children, mentally retarded children, and healthy examinees in qualitative and quantitative, but no differences in ranking the mean digital ridge count in autistic children. Sank and Sank-Firshein (24) did not find the differences in qualitative between autistic children from New Jersey and Florida and different population groups, but did detect differences between autistic and schizophrenic children. Arrieta et al (25) found differences between autistic male children and healthy control children, but not between autistic female children and healthy controls. These studies confirmed that dermatoglyphs of autistic children differed from healthy controls as well as from schizophrenic or mentally retarded children. The study of inheritance of s is based on the polygenetically determined laws (26). The sex-linked character has an effect on the values of the correlation coefficient. If a given character is X-linked, the mother-son and sister-sister correlation coefficients tend to be relatively higher (27). Figure 1. The right palm of an autistic male patient with severe damage and without normal palmar flexion creases. Table 1. Studies of dermatoglyphs in patients with autistic disorder Author (ref. No.) Subjects included in the study Statistical analysis Traits analyzed Finding indicating a strong genetic component Walker, 1977 (18) 78 autistic and 78 healthy children chi-square test qualitative "normal" palmar configurations between autistic and healthy children 8:1. Walker, 1977 (18) Hartin and Barry, 1979 (23) Hartin and Barry, 1979 (23) Sank and Sank Firshein, 1979 (24) 78 autistic and 78 healthy children 44 autistic, 44 mentally retarded, and 44 healthy control children 44 autistic, 44 mentally retarded, and 44 healthy control children 23 autistic children fromnew Jersey and 21 from Miami Sank and Sank Firshein, 44 autistic children and (24) schizophrenics Arrieta et al, 1990 (25) Arrieta et al, 1990 (25) 60 autistic Basque children and 71 healthy children 60 autistic Basque children and 71 healthy children t-test chi-square test Kolmogorov- Smirnov one-sample test chi-square test chi-square test quantitative qualitative quantitative qualitative qualitative G- test qualitative t-test quantitative observable diminishment of dermal ridge development arch patterns as a sensitive indicator for distinguishing autistic children no differences in ranking mean digital ridge count in autistic children compared with Walkers's study no differences between autistic children from different population groups differences between autistic children and schizophrenics differences between autistic and healthy control children found in boys but not in girls differences between autistic patients and healthy controls found in boys but not in girls 470

3 The aim of our study was to analyze quantitative in autistic patients and their families to confirm the differences found in previous studies between the affected and the healthy control groups. The main point of our study was to see whether there was a correlation in between autistic patients and their parents, brothers, and sisters, because no such collection of fingerprints of autistic patients and their family members has been made before. The data could provide further evidence for the confirmation of genetic transmission of autism. Subjects and Methods Subjects The sample included 120 autistic patients (92 male and 28 female patients aged between 1 and 30 years). We also analyzed all family members who consented to participate during the continual check-ups of their children. There were 92 mothers (74 of autistic male patients and 18 of autistic female patients) and 70 fathers (58 of autistic male patients and 12 of autistic female patients) aged between 25 and 70 years; and 28 sisters (22 of autistic male patients and 6 of autistic female patients) and 32 brothers (27 of autistic male patients and 5 of autistic female patients) aged between 1 and 30 years. In most families the autistic child was the only child, although there were some families with two autistic children. The control group comprised 400 phenotypically healthy unrelated controls from the population study carried out in the city of Zagreb (200 men and 200 women aged between 14 and 25). The group was fairly comparative to the autistic children in age, although this bears little importance because do not change with age (20). Two psychiatrists at the Diagnostic Department for Autism, Jankomir Psychiatric Hospital, Zagreb, Croatia, diagnosed autism following the criteria of DSM-IV (28) in patients hospitalized during the period. Informed consent was obtained from all participants or their parents. Assessment of Dermatoglyphs Finger and palm prints of both hands were collected and analyzed according to the Cummins and Midlo method (20). The number of ridges counted on each finger consisted of 10 variables (Fig. 2) of the finger ridge count (FRC): finger ridge count right for each finger on the right hand (FRCR1-FRCR5), and the finger ridge count left for each finger on the left hand (FRCL1- FRCL5). The counting was done along the straight line connecting the triradial point, formed by the confluence of the three ridge systems, to the point of core, which represents the approximate center of the pattern (20). In the case of palms (Fig. 3), the number of ridges was counted, according to those in the interdigital areas encompassing the following 6 variables: the a-b, b-c, and c-d ridge count (RC) on the right and left hand, between the strait line connecting two digital triradial points located on the base of each finger, ie, a, b, c, and d. The two atd angles were measured between a, t, and d triradial points on the right (R) and left (L) hand. Following this procedure, the atd angle (measured in degrees) was formed by drawing lines from the digital triradius a to the axial triradius t (at the base of the palm) and to the digital triradius d. Due to high correlations between respective on the right and left hand of the same person, the number of variables was reduced by the addition of the ridge counts of respective, ie, instead of both FRCR1 and FRCL1, their sum FRC1=FRCR1+FRCL1 was taken for further analysis, as well as respective sums FRC2-FRC5 for fingers and a-b, b-c, and c-d ridge counts (RC) on the palm. The two atd angles were considered separately. The number of variables was thus reduced to 10: five fingers (FRC1-FRC5), three palm (,, and c-d RC), and two atd angles ( and ). Statistical Analysis The analysis encompassed descriptive statistics (means± standard deviations), multivariate analysis of variance (MANOVA) with Tukey HSD post hoc test for multiple comparisons and discriminant analysis for eight groups of examinees (29,30). Interclass correlation coefficients were calculated for paired familial data (31). P-values less than 0.05 were considered statistically significant. Figure 3. A print of the right hand palm with all counting elements. On the palm, print ridges were counted between the straight line connecting two digital triradii located on the base of the each finger: a, b, c, and d. The atd angle was formed drawing the lines from the digital triradius a to the axial triradius t (on the base of the palm) and to the digital triradius d, and it was measured in degrees. Three flexion creases (distal, proximal, and thenar), were indicated by arrows only for comparison with Fig 1. Figure 2. A print of the right hand fingers with all counting elements. For each finger, the counting was done along the straight line connecting the triradial point (a triradius is formed by the confluence of the three ridge systems) to the point of the core (the core is the approximate center of the pattern). The number of ridges was presented as the finger ridge count right (FRCR), and the fingers were numbered from the thumb (FRCR1) to the little finger (FRCR5). In the case of whirls, the higher ridge count was taken (FRCR2-FRCR5). 471

4 Results Differences We analyzed all 18 variables, presented as mean±sd separately for male (Table 2) and female participants (Table 3). Ten variables (FRC1-FRC5,,,,, and ) were entered into factorial multivariate analysis of variance (MANOVA) with two factors, sex (male/female) and four groups regarding autism (autistic patients, father/mother of an autistic patient, brother/sister, and healthy control), giving altogether eight groups. Both factors showed to be significant among the examined groups. The differences among the Table 2. Dermatoglyphic variables (mean±sd)ofthedigito- palmar complex in autistic male patients, their fathers, brothers, and healthy male controls Autistic male patients (n=92) Fathers (n=70) Healthy brothers (n=32) Healthy male controls (n=200) Variables* Right hand: FRCR1 17.5± ± ± ±5.6 FRCR2 10.3± ± ± ±7.3 FRCR3 11.0± ± ± ±6.6 FRCR4 13.5± ± ± ±6.2 FRCR5 11.7± ± ± ±5.2 R 38.9± ± ± ±6.9 R 27.5± ± ± ±5.8 R 36.0± ± ± ± ± ± ± ±8.3 Left hand: FRCL1 14.5± ± ± ±6.1 FRCL2 10.4± ± ± ±6.8 FRCL3 11.2± ± ± ±6.4 FRCL4 13.5± ± ± ±6.2 FRCL5 11.9± ± ± ±4.6 L 40.1± ± ± ±7.1 L 26.1± ± ± ±5.7 L 34.5± ± ± ± ± ± ± ±7.7 *Abbreviations: FRCR finger ridge count right; FRCL finger ridge count left; R, R, and R palmar ridge count right; L, L, and L palmar ridge count left; right angle; and left angle. Table 3. Dermatoglyphic variables (mean±sd)ofthedigito- palmar complex in autistic female patients, their mothers, sisters, and healthy female controls Autistic female patients (n=28) Mothers (n=92) Healthy sisters (n=28) Healthy female controls (n=200) Variables* Right hand: FRCR1 15.0± ± ± ±5.6 FRCR2 9.8± ± ± ±6.6 FRCR3 11.2± ± ± ±5.3 FRCR4 12.2± ± ± ±5.7 FRCR5 10.3± ± ± ±4.8 R 39.1± ± ± ±6.0 R 27.2± ± ± ±5.9 R 36.0± ± ± ± ± ± ± ±8.7 Left hand: FRCL1 13.8± ± ± ±5.8 FRCL2 9.6± ± ± ±6.9 FRCL3 11.2± ± ± ±5.7 FRCL4 13.1± ± ± ±5.3 FRCL5 10.7± ± ± ±4.8 L 41.0± ± ± ±5.9 L 25.6± ± ± ±5.5 L 33.6± ± ± ± ± ± ± ±8.4 *Abbreviations: FRCR finger ridge count right; FRCL finger ridge count left; R, R, and R palmar ridge count right; L, L, and L palmar ridge count left; right angle; adt L left angle. groups were significant (p<0.05) for all variables except FRC2 and FRC3 (Table 4). Post hoc multiple comparisons with Tukey HSD test were made to reveal statistically significant pairs of groups (Table 5). Male and female controls differed only in a single variable, FRC1. This fact, together with the appearance of FRC1 in other pairs of male-female groups (mothers vs fathers, male controls vs mothers, and male controls vs female autistic patients) strongly suggested that FRC1 was influenced by sex. Female controls differed from mothers of autistic patients in FRC1, FRC4, FRC5,,, and both atd angles, whereas male controls differed from fathers of autistic patients only in two atd angles. The differences between mothers of autistic patients and female controls were more pronounced than the differences between fathers and male controls. Male controls differed from male autistic patients in the fourth and fifth finger, and both atd angles and from brothers of autistic patients in all palmar variables but in none of the digital variables. The fourth finger made a difference between male controls and members of affected families, especially female members: it separated male controls from male autistic patients and from all three female groups of family members (female autistic patients, mothers, and sisters). Generally speaking, it seemed that the differences among members of affected families and healthy controls were pronounced in palmar variables and in the fourth and fifth finger. The results of multivariate discriminant analysis for eight examined groups and 10 variables revealed that out of seven canonical discriminant functions, the first two were significant at p<0.05 level, and that cumulative percentage of variance explained by those two functions was high (85.4%). Structure matrix gives the correlations between original variables and discriminant functions (variables were sorted by ascending absolute correlation between each variable and any discriminant function; Table 6). Standardized canonical discriminant function coefficients are given in Table 7. Table 8 shows coordinates of group centroids in discriminant space. Discriminant function 2, which had the highest correlation with variable FRC1, clearly separated male from female participants, both in healthy controls and family members, whereas discriminant function 1, which correlated with atd angles of left and right hand, a-b, c-d, and FRC4, separated healthy controls from family members (Fig. 4). When group centroids on discriminant functions 3 and 4 were graphically presented, the discriminant function 3 (defined mostly by ) had autistic male patients and healthy brothers on its opposite poles (fathers were somewhere in between and mothers were on the side of autistic male patients), whereas the discriminant function 4, which correlated with FR4 and, separated autistic female patients from all other groups, particularly from mothers and female controls, who were at opposite ends (sisters were somewhere in between) (Fig. 5). 472

5 Similarities To search for the similarities between family members, we correlated 10 variables for all pairs of available family members (fathers, mothers, and affected and healthy children; Table 9). Statistically significant correlation was found between the mothers of autistic sons and their sons in 7 out of 10 variables (except for and both atd angles). The fathers also correlated with their autistic sons in 7 out of 10 variables, but did not correlate for the second and fifth finger and for atd angle of the right hand. However, the correlation between parents and their autistic daughters was lower (three and two variables for mothers and fathers, respectively), as well as between parents of autistic patients and their healthy sons and daughters. Interestingly, both the mothers and the fathers correlated with their healthy children in palmar variables only. The similarity of both the mothers and the fathers with their autistic sons was very high, but it was low with either their autistic daughters or healthy sons and daughters. The correlation between mothers and their healthy sons was significant for a single variable only (c-d). Discussion Our study showed that autistic male patients, their fathers, mothers, and brothers expressed signifi- Table 4. Multivariate analysis of variance (MANOVA) for 10 variables; factorial model for sex (male/female) and group membership regarding autism (four groups: autistic patient, father/mother of an autistic patient, brother/sister of an autistic patient, and healthy control) Whole model Hypothesis decomposition mean square p-values by source Variable* model (df=7) error (df=734) F p sex (df=1) group (df=3) sex group (df=3) FRC <0.001 < FRC FRC FRC < < FRC <0.001 <0.001 < < < < < < < < < < < *Abbreviations: FRC finger ridge count;,, and palmar ridge counts; right angle; and adt L left angle. Table 5. Multiple comparisons for 8 groups of participants and significant probabilities (p<0.05) according to post hoc Tukey HSD tests* Male participants Female participants autistic fathers healthy brothers controls autistic mothers healthy sisters controls (n=92) (n=70) (n=32) (n=200) (n=28) (n=92) (n=28) (n=200) Male participants autistic FRC4 FRC5 fathers FRC1 FRC1 FRC5 healthy brothers controls FRC1 FRC1 FRC4 FRC1 FRC4 FRC4 FRC5 FRC5 Female participants autistic mothers FRC1 FRC4 FRC5 healthy sisters controls *Abreviations: FRC finger ridge count;,, and palmar ridge count; right angle; and left angle. 473

6 Table 6. Discriminant analysis: structure matrix showing correlations among original variables and canonical discriminant functions (upper part of the table), and eigenvalues and chi-square tests for 7 canonical discriminant functions discriminating 8 examined groups (lower part of the table). First two discriminant functions were significant explaining 85.4% of variance. Discriminant function 1 correlated with atd angles of both hands, a-b and, FRC4, and FRC5, whereas discriminant function 2 had the highest correlation with FRC1 Discriminant function Variable* FRC FRC FRC FRC FRC Eigenvalues Cumulative % of variance chi-square tests when preceding roots were removed p (chi-square test) < *Variables are ordered by absolute size of correlation within function (largest absolute correlation between each variable and any discriminant function). Abbreviations: FRC finger ridge count;,, and palmar ridge count; right angle; and adt L left angle. Table 7. Standardized canonical discriminant function coefficients, enabling transformation of standardized values of original variables into scores on discriminant functions Discriminant function Variable* FRC FRC FRC FRC FRC *Abbreviations: FRC finger ridge count;,, and palmar ridge count; right angle; and adt L left angle. Table 8. Discriminant functions at group centroids. Discriminant functions at group centroids: discriminant function 1 separates healthy controls from family members, whereas discriminant function 2 clearly separates male participants from female ones for both healthy controls and family members Discriminant function Group Autistic male patients Autistic female patients Fathers Mothers Healthy brothers Healthy sisters Male controls Female controls cant differences in the number of ridges in as compared with the healthy control groups. These differences were not so evident (lower ridge counts existed but were not statistically significant) in the group of autistic daughters and their families, but MANOVA and discriminant analysis clearly separated all examined patients (male and female) and their family members from the control groups. A rather small sample size could be one of the reasons for the lower number of differences found between autistic female patients, their mothers, and their sisters when compared with the healthy female controls. These results are in concordance with several other studies on dermatoglyphs of the digito-palmar complex in autistic patients. We also analyzed interclass correlations of the variables among family members, and found them to be more numerous between mothers and their autistic sons than between mothers and their autistic daughters. Statistically significant differences separating the tested groups with respect to sex, and separating the group of healthy controls from both the autistic male patients and their family members (father, mother, brother, and sister), on the one side, and correlations in greater number of variables found between mothers and their autistic sons, on the other, indicated that the parents of autistic children could also be carriers of certain genetic modifications shown in the dermatoglyphs. We found the results intriguing because dermatoglyphs of the digito-palmar complex are polygenetically determined and the correlation between family members is well known (21). High number of correlated variables among affected family members and large number of differences from the healthy controls could indicate the genetic transition of some diseases in the families. Our results are in line with the fact that the timing of the early stages of development of the neurological system is simultaneous to that of the dermal ridges (12). Any damage occurring at that time can harm the embryonic development processes and later on be expressed through both systems, dermal (expressed in dermatoglyphs) and neurological (expressed as autism). If the occurrence of variables with autism is not accidental, the transition of the de- 474

7 Discriminant function 2 Discriminant function Brothers Mothers Autistic males Autistic males Control males 0 Mothers Sisters Fathers Autistic females Sisters Control females Fathers Brothers Discriminant function 3 Control females Control males Autistic females Discriminant function 1 Figure 4. Graphical presentation of group centroids on discriminant functions 1 and 2. Discriminant function 1 (horizontal) clearly separated healthy controls (right) from family members (left), whereas function 2 (vertical) separated male participants (positive coordinates of group centroids) from female participants (negative coordinates). Figure 5. Graphical presentation of group centroids on discriminant functions 3 and 4. Discriminant function 3 (horizontal) defined mostly by b-c ridge count (RC) has autistic male patients and healthy brothers on its opposite poles (negative and positive, respectively); function 4 that correlated with finger ridge count 4 (FRC4) and left atd angle () separated autistic female patients from all other groups. Table 9. Statistically significant interclass correlations of familial data (classical pair-wise test for familial data were used for 10 dermatogliphyc variables)* Children Male Female autistic healthy autistic healthy Parents patients brothers patients sisters Mothers n=70 n=30 n=14 n=26 FRC1 FRC3 FRC2 FRC5 FRC3 FRC4 FRC5 Fathers n=55 n=24 n=12 n=25 FRC1 FRC2 FRC3 FRC3 FRC4 *Abbreviations: FRC finger ridge count;,, and palmar ridge count; right angle; left angle; only statistically significant correlations coefficients at p<0.05 are shown. fect on the X-chromosome may more often affect male patients, because in female patients one of the affected X-chromosomes may be inactivated. Analyzing the X-chromosome inactivation detected in 14.4% healthy female participants, linkage analysis in phenotypically normal families showed that a linkage on the loci Xq25 Xq26 was very close to the fragile X loci (32). Hogben (27) has shown that, when a given character is X-linked, mother-son and sister-sister correlation coefficients tend to be relatively higher than within the other family members. The frequency of autistic disorders is 3 to 4 times higher in boys than in girls (25). Studies on families ascertained through a single autistic proband suggest that higher rates of social and communication defects and stereotyped behaviors exist in relatives of the families with a multiple incidence of autism (33). The recurrence risk estimation (a chance that each sibling born after an autistic child develops autism) in male children is 7% and in female children 14.5% (6). The incidence of the fragile X in autistic patients ranges from 0% (1) to 12% or even higher (15,16). Translocation X-8 was identified in female autistic patients (34). Using the multipoint sib-pair linkage analysis on 35 microsatellite markers located on the X-chromosome, Hallmayer et al (35) excluded any moderately strong effect causing autism on the X-chromosome, but concluded that smaller gene effects (lambda sub XS<4) could be located between DXS453 and DXS1001. In family studies, in cases of autism associated with fragile X chromosome, three folate-sensitive fragile sites in the Xq27 <Xq28 region were found (36,37). Genes on the X chromosome are important in nonspecific mental retardation (38). Our study showed that healthy controls differ in not only from autistic patients, but also from their phenotypically healthy family members. On the other hand, we found much greater similarity of mothers and fathers with their autistic sons than with their autistic daughters or healthy children. Our results point to the existence of genetic changes in relatives, mostly in parents of autistic patients, and confirm the hypothesis of Folstein and Rutter (33) that a subset of autism might be caused by a gene located on the X chromosome, whereas other cases are caused by a combination of autosomal genes, environmental influences, or their interactions. Acknowledgment This work was sponsored by the Ministry of Science and Technology of Republic of Croatia Grant No Special thanks go to prof. Mirna Jemriæ Kovaèiæek and colleagues Drs Tatjana Škariæ-Juriæ and Sanja Špoljar-Vr ina for their valuable help. References 1 Smalley SL, Asarnow RF, Spence MA. Autism and genetics. A decade of research. Arch Gen Psychiatry 1988;45: Wassink TH, Piven J. The molecular genetics of autism. Curr Psychiatry Rep 2000;2:

8 3 Rutter M, Bolton P, Harrington R, Le Couteur A, Mac- Donald H, Simonoff E. Genetic factors in child psychiatric disorders - I. A review of research strategies. J Child Psychol Psychiatry 1990;31: Rapin I. Autism. N Engl J Med 1997;337: Folstein S, Rutter M. Infantile autism: a genetic study of 21 twin pairs. J Child Psychol Psychiatry 1977;18: Ritvo ER, Jorde LB, Mason-Brothers A, Freeman BJ, Pingree C, Jones MB, et al. The UCLA University of Utah epidemiological survey of autism: recurrence risk estimates and genetic consulting. Am J Psychiatry 1989; 146: Hunt A, Dennis J. Psychiatric disorder among children with tuberous sclerosis. Dev Med Child Neurol 1987; 29: Hackney IM, Hanley WB, Davidson W, Lindsao L. Phenylketonuria: mental development, behavior, and termination of low phenylalanine diet. J Pediatr 1968;72: Gillberg C, Forsell C. Childhood psychosis and neurofibromatosis more than a coincidence? J Autism Dev Disord 1984;14: Bauman ML. Brief report: neuroanatomic observations of the brain in pervasive developmental disorders. J Autism Dev Disord 1996;26: Kimura S. Embriologic development of flexion creases. Birth Defects Orig Artic Ser 1991;27: Lainhart JE. Developmental abnormalities in autism. The Lancet 1997;349: Cook EH Jr, Courchesne R, Lord C, Cox NJ, Yan S, Lincoln A, et al. Evidence of linkage between the serotonin transporter and autistic disorder. Mol Psychiatry 1997; 2: Persico AM, Militerni R, Bravaccio C, Schneider C, Melmed R, Conciatori M, et al. Lack of association between serotonin transporter gene promoter variants and autistic disorder in two ethnically distinct samples. Am J Med Genet 2000;96: Gillberg C. Subgroups in autism: are there behavioral phenotypes typical of underlying medical conditions? J Intellect Disabil Res 1992;36(Pt 3): Reiss AL, Freund L. Behavioral phenotype of fragile X syndrome: DSM-III-R autistic behavior in male children. Am J Med Genet 1992;43: Bregman JD, Leckman JF, Ort SI. Fragile X syndrome: genetic predisposition to psychopathology. J Autism Dev Disord 1988;18: Walker H. Dermatoglyphic study of autistic patients. J Autism Child Schizophr 1977;7: Schaumann B. Dermatoglyphics in medical disorders: an overview. Coll Antropol 1987;11: Cummins H, Midlo C. Finger prints, palms and soles. 2nd rev. ed. New York (NY): Dover Publications Inc.; Loesch D. Genetical studies of palmar and sole patterns and some measurements in twins. Ann Hum Genet 1979;43: Hauser G. Dermatoglyphic today. Coll Antropol 1987; 11: Hartin PJ, Barry RJ. A comparative study of autistic, retarded, and normal children. J Autism Dev Disord 1979;9: Sank D, Firschein BD. Fingerprints and laterality preferences of early-onset autism. Birth Defects Orig Artic Ser 1979;15: Arrieta MI, Martinez B, Criado B, Simon A, Salazar L, Lostao CM. Dermatoglyphic analysis of autistic Basque children. Am J Med Genet 1990;35: Fisher RA. Statistical methods for research workers. 10th rev. ed. Edinburgh: Oliver and Boyd; Hogben L. Filial and fraternal correlations in sex-linked inheritance [retracted in Loesch DZ. Quantitative s classification, genetics, and pathology. Oxford: Oxford University Press; 1983]. Proc R Soc Edinb Sect B (Biol) 1932;52: American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 4th ed. Washington (DC): American Psychiatric Association; SAS Institute. The SAS System for Windows [computer program]. Version Cary (NC): SAS Institute Inc; StatSoft, Inc. Statistica [computer program]. Version 6.0. Tulsa (OK): StatSoft; Rosner B, Donner A, Hennekens CH. Significance testing of interclass correlations from familial data. Biometrics 1979;35: Naumova AK, Olien L, Bird LM, Smith M, Verner AE, Leppert M, et al. Genetic mapping of X-linked loci involved in skewing of X-chromosome inactivation in the human. Eur J Hum Genet 1998;6: Folstein S, Rutter M. Autism: familial aggregation and genetic implications. J Autism Dev Disord 1988;18: Ishikawa-Brush Y, Powell JF, Bolton P, Miller AP, Francis F, Willard HF, et al. Autism and multiple exostoses associated with an X-8 translocation occurring within the GRPR gene and 3 to SDC2 gene. Hum Mol Genet 1997;6: Hallmayer J, Hebert JM, Spiker D, Lotspeich L, McMahon WM, Petersen PB, et al. Autism and the X chromosome. Multipoint sib-pair analysis. Arch Gen Psychiatry 1996;53: Warren ST, Nelson DL. Advances in molecular analysis of fragile X syndrome. JAMA 1994;271: Holden JJ, Wing M, Chalifoux M, Julien-Inalsingh C, Schutz C, Robinson P, et al. Lack of expansion of triplet repeats in the FMR1, FRAXE, and FRAXF loci in male multiplex families with autism and pervasive development disorders. Am J Med Genet 1996;64: Partington M, Mowat D, Einfeld S, Tonge B, Turner G. Genes on the X chromosome are important in undiagnosed mental retardation. Am J Med Genet 2000;92: Received: November 2, 2001 Accepted: March 24, 2003 Correspondence to: Jasna Milièiæ Institute for Anthropological Research Amruševa 8 (P.O. Box 290) Zagreb, Croatia Jasna@luka.inantro.hr 476