Annotated Bibliography Project

Is season of birth related to developmental dyslexia?

Renato Donfrancesco & Roberto Iozzino & Barbara Caruso & Laura Ferrante & Daniele Mugnaini & Alessandra Talamo & Silvia Miano & Andrea Dimitri & Gabriele Masi

Received: 7 August 2009 /Accepted: 11 May 2010 /Published online: 3 August 2010 # The International Dyslexia Association 2010

Abstract Different moderators/mediators of risk are involved in developmental dyslexia (DD), but data are inconsistent. We explored the prevalence of season of birth and its association with gender and age of school entry in an Italian sample of dyslexic children compared to an Italian normal control group. The clinical sample included 498 children (345 boys, mean age 10.3±2.1 years) with DD, the control sample 1,276 children (658 boys, mean age 10.8±2.2 years) from four elementary schools from the same urban area, and with the same socio-economic status level. A prevalence of birth in autumn was found among children with DD compared to controls (34% versus 24%, p<0.0001). Children with DD were more frequently males (p<0.0001) and had a lower mean age of school entry (p< 0.0001). Regarding the distribution of ages, 11.4% of children with DD, but none of the subjects in the control group, started school before 5.7 years. Therefore, greater risk of DD

Ann. of Dyslexia (2010) 60:175–182 DOI 10.1007/s11881-010-0037-6

R. Donfrancesco : B. Caruso : L. Ferrante Child Neuropsychiatry Department, La Scarpetta Hospital, ASL RM/A, Rome, Italy

R. Iozzino Center for learning disorders, ASL RM/A, Rome, Italy

D. Mugnaini A. Meyer Children’s Hospital, Florence, Italy

A. Talamo Department of Neurosciences, Mental Health, and Sensory Functions (NESMOS), Second Medical School, Sapienza University of Rome, Sant’Andrea Hospital, Rome, Italy

S. Miano Department of Paediatric, Sleep Disease Centre, University of Rome La Sapienza-S.Andrea Hospital, Rome, Italy

A. Dimitri Nestor Lab, University of Rome Tor Vergata, Rome, Italy

G. Masi (*) Scientific Institute of Child Neurology and Psychiatry, IRCCS Stella Maris, Via dei Giacinti 2, 56018 Calambrone, Pisa, Italy e-mail: [email protected]

was related to age of school entry (OR=2.72), gender (OR=2.16), and season of birth (OR =1.21). Significant interactions between boys with DD born in autumn, and correct school of entry (OR=2.56) were joint predictors of higher risk of DD. The association between birth in autumn and DD may be explained by the earlier age of school entry, which may be a critical element in the youngest children with DD or at risk to DD. Whether Italian school policy is oriented to anticipate the school entry, a closer detection of early learning disorders and associated risk factors (familial load, specific language disorders, and attention deficit hyperactivity disorder) should be warranted.

Keywords Academic achievement . Children . Cognitive development . Learning disabilities . Prevention

Introduction

Developmental dyslexia (DD), or specific reading disability, is the most common learning disability, accounting for about 80% of all learning disabilities cases (Lyon, 1996; Shaywitz, 1998). It affects approximately 4–12% of children in languages characterised by non-transparent orthography, such as English, and 3–8% of children, in those characterised by strict grapheme–phoneme correspondence, such as Italian (Lindgren, De Renzi, & Richman 1985). Different causal mechanisms can be involved in DD, including anatomical, genetic and environmental factors (Grigorenko, 2001; Shaywitz & Shaywitz, 2005; Williams & O'Donovan, 2006). The month/season of birth (Flynn, Rahbar, & Goodman 1996; Geldhill, Ford, & Goodman 2002), gender (Shaywitz, Shaywitz, Fletcher, & Escobar 1990; St Sauver, Katusic, Barbaresi, Colligan, & Jacobsen 2001; Liederman, Kantrowitz, & Flannery 2005), and age of school entry (Flynn et al., 1996; Lawlor, Clark, Ronalds, & Leon 2006) have been proposed as possible mediators of risk, but data are still inconsistent.

The role of season of birth has been repeatedly reported in mental and neuro- developmental disorders, including DD (Geldhill et al., 2002). The mechanisms underlying this association are still unknown, although ambient temperature, viral infections, maternal hormones, melatonin exposure, sperm quality, and external toxins have been proposed as mediators. An interaction between season of birth and genes has also been reported (Levitan et al., 2006). All previous studies (except for Flynn’s study), did not involve a sample of children affected by DD, but only epidemiological samples of normal children. The aim of our study is to explore the prevalence of season of birth among a large sample of Italian children with a diagnosis of DD, compared to a control group, and to evaluate its association with gender and age of school entry. We supposed that age of school entry may be an indicator of poor early education in Italian children at risk for DD.

Patients and methods

Sample

Among the outpatient children assessed in the period September 1999-September 2006, in the Center for Developmental Language and Learning Disorders of the Rome Public Health Service, a secondary care unit which received a large number of patient from Roman urban area (Italy), 498 Caucasian subjects (345 boys and 153 females, mean age of 10.3± 2.1 years) received a diagnosis of DD, according to Diagnostic and Statistical Manual of

176 R. Donfrancesco et al.

Mental Disorders, Fourth Edition criteria (American Psychiatric Association, 1994). These criteria include normal intelligence quotient (IQ) and impaired reading performance, both assessed with standardised measures. All the children were free from neurological, sensory, and/or educational deficits. All subjects were in the middle to upper middle socioeconomic status level, with a mean score of 59.5±10.5 according to Hollingshead two-factor model (Hollingshead, 1957). All were from second to eight grade of primary school.

The gender, month of birth and the age of school entry were collected, and compared to those of a control group of 1,276 Caucasian children (658 boys and 618 girls, mean age of 10.8±2.2 years), collected in July 2006, from all the registers of four primary schools. The data of controls were collected from four out of 20 schools from the same urban area, in order to obtain the same socio-economic status level (the socio-economic status level score according to Hollingshead two-factor model was 59.3±7.4). The schools were numbered from 1 to 20 and then four were chosen by a selection from a random number table. Even the control children were from second to eight grade of primary school. None of the controls recruited for the study, had difficulty reading, evidence of cognitive impairment, attention deficit, or a hyperactivity disorder, reported by parents or by teachers.

We did not exclude any Italian and Caucasian students from the study in order to obtain a large sample of children attending primary school with the same standard Italian teaching protocol of children with DD. After we did include children without DD, we excluded children who were not Caucasian and were members of a non-native Italian family. The age of school entry is different from the age of other country, since in Italy elementary school starts at September.

Measures

Criteria for inclusion in the group with DD were a normal IQ, assessed with a standardised measure, the Wechsler Intelligence Scale for Children Revised (Wechsler, 1986), and a significant reading delay, with a score below at least 2 SD, in terms of speed and accuracy, by means of a standardised battery for the diagnosis of DD in Italian children (Job, Sartori, & Tressoldi 1995; Tressoldi, Stella, & Faggellaet 2001). This test is composed of five subtests for the assessment of reading, and three subtests for the assessment of writing. The subtests for reading include a single grapheme identification, a lexical decision task, four lists of isolated words of different frequency, three lists of non-words of different orthographic complexity, and three subtests for the identification of homophones. Norms are drawn from different samples from the second to the eighth grade in a large sample of Italian children, representative of Italian general population. The lists of isolated words were specifically used for diagnosis of DD in our children. These subtests are similar to the Reading (Word Recognition) subtests of the Wide Range Achievement Test–Revised (Jastak & Wilkinson, 1984). Parallel and construct validity evidence are psychometrically appropriate and are reported in the manual (test-retest values 0.91).

Data analyses

We used chi-square test of independence to investigate the relationship between gender and dyslexia; season of birth and dyslexia; season of birth were defined as: winter (December, January, or February); spring (March, April, or May); summer (June, July, or August); and fall (September, October, November). The chi-square test was also used to assess relationship between age of school entry (early versus overage school of entry, mean±

Is season of birth related to developmental dyslexia? 177

SD) and dyslexia. The ANOVA test was used on continuous variables with Bonferroni post hoc correction, setting significance at 0.002 level, two-tailed.

Finally, we used logistic regression to investigate the joint effect of gender, age of school entrance, and season of birth on presence of dyslexia. Similar to the multiple regression model for continuous variable, logistic regression models of a dichotomous dependent variable, in this case presence or absence of dyslexia, in a logistic regression model, the probability (p) of presence is represented by:

P! exp^ B0"B1 gender# $"B2 age school entry# $"B3 birth season# $% &=

1 " exp^ B0"B1 gender# $"B2 age school entry# $"B3 birth season# $% & n o

Where B1 is the natural log of the odds for that factor (gender adjusted for the other factors in the mode; age category and season of birth) this model results in adjusted coefficients that are used to estimate adjusted odds ratios, i.e., the odds of experiencing a given outcome for one factor adjusted for the presence and influence of the other factors in the model.

Statistical analyses were based on commercial software (Statview-5®, SAS Corporation, Cary, NC, USA; Stata-8,® Stata Corporation, College Station, TX, USA).

Results

Among 1,774 children of the combined research sample, 28.07% (N=498) had dyslexia. The effect size for boys number was 1.085 (power=0.99), for early overage was 1.45 (power 0.99), for season of birth 0.49 (power 0.99), for age was 0.61 (power 0.99). A higher percentage of boys (69% vs. 52%, p<0.0001) and a lower mean age of school entry (6.0±.4 years vs. 6.2±.3 years, p<0.0001) were found in patients with DD compared to the control group (Table 1). A different distribution of season of birth was found between the groups (p<0.0001), with a prevalence of children with DD born in autumn compared to controls (34% versus 24%, chi-square=20.93, p<0.0001; Table 1). No statistical differences appeared in the mean age at school entry between gender. Boys with DD had a mean age of school entry 6.0±.4 years, girls with DD 6.0±.3; boys from the control group 6.2±.3 years, girls from the control group 6.2±.3 years.

The age range of school entry was 5.0-6.8 years for children with DD and 5.9-6.8 years for controls. An analysis of the distribution of the ages revealed that 11.4% of children with DD, but none of the subjects in the control group, started school before 5.7 years. Among the overall sample of correct age of school entrance (N=1,559), 25.40% (N=392) had dyslexia, one out of four, and among children with earlier school of entry (N=215), 49.30% (N=106) had dyslexia, one out of two.

Multivariate analysis by logistic regression, following preliminary findings of bivariate comparisons, involving 1,774 subjects (N=498 children with dyslexia versus N=1,276 controls) were considered. Table 2 shows odds ratios for DD based on gender, season of birth, and age of school entry (model 1). We found greater risk of DD related to age of school entry (OR=2.72), male gender (OR=2.16), and season of birth (OR=1.21). Based on the above results, we also studied interactions between season of birth and gender considering the following logistic regression model (Table 2; model 2):

P! exp^ B0"B1 interaction# $% &= 1 " exp^ B0"B1 interaction# $% &# $:

178 R. Donfrancesco et al.

We found a significant model of interactions between boys with DD born in autumn and correct school of entry (OR=2.56), which were jointly predictors of higher risk of DD.

Discussion

To our knowledge, this study is the first attempt to demonstrate an association between DD and gender, season of birth, and age at school entry in a large Italian clinical sample of

Table 2 Odds ratios for dyslexia based on gender, season of birth, age of school entry, and the interaction between season of birth, gender and age of school entry

Independent variables B Standard error p Odds ratio

Model 1

Gender 0.7710 0.1143 <0.0001 2.16

Age of school entry 1.0034 0.1339 <0.0001 2.72

Season of birth 0.1945 0.0532 0.0003 1.21

Model 2

Interactiona 0.9300 0.141 <0.0001 2.56

Multivariate analysis is by logistic regression, following preliminary findings of bivariate comparisons, involving 1,774 subjects (N=498 children with dyslexia versus N=1,276 controls). Model !2 =117.31, df=3, p<0.0001 (model 1) a Interaction considers boys with dyslexia born in autumn, and correct school of entry (P=exp^[B0+B1 (interaction)])/(1+ exp^[B0+B1(interaction)]). Model !2 =43.02, df=1, p<0. 0,001 (model 2)

Table 1 Characteristics of subjects with dyslexia compared to controls

Subjects with dyslexia (n=498)

Controls (n=1,276)

Chi- square/F

DF p< Post hoc Bonferroni p<

Boys 345 (69.3%) 658 (51.5%) 45.7 1 0.0001

Girls 153 (30.7%) 618 (48.5%)

Age

Age at school entrance 6.02±0.4 6.19±0.3 108.6 0.0001 0.0001

Years, mean±SD 6.01±0.4 6.20±0.3 108.6 0.0001 0.0001

Months mean±SD 72.33±4.5 74.35±3.3 108.6 0.0001

School entrancea

Early 106 (21.3%) 109 (8.5%) 61.20 1 0.0001

Overage 392 (78.7%) 1,167 (91.5%)

Season of birth 20.93 3 0.0001

Autumn N=169 (33.9%) N=310 (24.4%)

Winter N=107 (21.5%) N=267 (20.9%)

Spring N=107 (21.5%) N=370 (28.9%)

Summer N=115 (23.1%) N=329 (25.8%)

a Overall mean age of school entry (N=1,774 age months; 73.78±3.78); early=mean SD [<70 months]; overage=mean+SD [!70 months]. DF were for chi-square test.

Is season of birth related to developmental dyslexia? 179

children with DD. A higher percentage of boys (69% vs. 52%) and a lower mean age of school entry (6.0±0.4 years vs. 6.2±0.3 years) were found in patients with DD compared to the control group (Table 1). Among children with correct age of school entrance, one out of four versus one out of two with earlier school of entry had dyslexia.

This result showed a 200% greater rate of DD among children with early school of entry. Interestingly, we also found that boys born during autumn and correct school of entry had almost a threefold risk of presenting a DD. The preponderance of male gender, still controversial and probably weak, is supported by other studies (Liederman et al., 2005). Although a significant difference was found between dyslexic and control children according to the age at school entry (about 2 months), this difference was slight (about 2 months), but logistic regression did show a significant association between DD, date of school entry, and gender. More significantly, about 11% of children with DD were aged less than 5.7 years at their school entry, compared to none in the control group.

Other studies have explored the role of season of birth in children with DD. In a mixed population of boys with different neurodevelopmental disorders, including DD, Livingston, Adam, & Bracha (1993) found a greater incidence of DD for births in May, June, and July, and hypothesised that viral infection (influenza) in the second and third trimester of pregnancy may have been an explanation of this finding. Flynn et al. (1996) explored the role of season of birth, age at school entrance, and reading failure in two cohorts of children (n=2,411 and n=1,972). Logistic regression showed a significant interaction between reading failure and overage at school entry, which was caused by kindergarten entrance cut-off birth dates. This study examined the role of school entry and season of birth in two larger cohorts of North American second grade children from urban and rural community and can be considered an epidemiological study for DD. The age of entry was different from Italian age, since North American school entry was December. In our study, we compared a large cohort of outpatients from a secondary care unit for the diagnosis of DD, with a control group from the same urban area, and we should take in consideration that children were selected from a secondary care unit, and not from an epidemiological sample. For all of this reasons, our results cannot be compared with those of Flynn et al. (1996), even if the same authors found that age at school entry is a bias to the interpretation that season birth may predict DD in children. The role of age of school entry is supported by a study investigating the development of learning abilities in a cohort of 12,150 individuals born in Aberdeen (Scotland) between 1950 and 1956, assessed at the age of 7, 9, and 11, by season of birth (Lawlor et al., 2006). The authors found that child’s perception and understanding of pictorial differences at age7 and reading ability at age9 were both affected by age of school entry, with a low scores in these tests in children who had a school age of entry younger than 61–63 months (the most common age range for school entry) and whose within the upper range of school entry. Moreover, the same authors found that reading ability at age 9 and arithmetic ability at age11 were lowest among children born in autumn (September–December), while the highest scores were found in those born in later winter or early spring (February-April). However, these differences were small and were further reduced towards the null after adjustment for age at school entry and age relative to class peers. The role of earlier school entry as a possible risk factor for of emotional and behavioural problems is supported by a cross-sectional survey including 10,438 participants (5-15 year olds) from England, Scotland, and Wales (Goodman, Gledhill, & Ford 2003). According to this study, younger children in a school year were significantly more likely to have higher symptom scores and psychiatric disorder. Unfortunately, we

180 R. Donfrancesco et al.

did not provide data about psychiatric comorbidity in our sample of children with DD, and this may limit the interpretation of our results.

Our findings suggest that although the role of possible risk factors (temperature, infections, etc) cannot be ruled out, the strong association between birth and DD may be simply explained by the association with earlier age of school entry and on the other hand associated to gender differences. Whether age of school entry is a major element in developing DD, implications may be relevant in educational systems with a single annual cut-off date for school entry, when the age of children can vary within a 12-month range. In the youngest children with DD or at risk to DD, the earlier age of starting school may be a critical element in the expression of the disorder. Whether school policy is oriented to anticipate the school entry, a closer detection of early learning disorders and associated risk factors, such as familial load, specific language disorders, and/or attention deficit hyperactivity disorder should be warranted for prevention and/or timely treatment of these disorders at their first presentation.

References

American Psychiatric Association. (1994). Diagnostic and Statistical Manual of Mental Disorders (4th ed.). Washington: American Psychiatric Association.

Flynn, J. M., Rahbar, M. H., & Bernstein, A. J. (1996). Is there an association between season of birth and developmental dyslexia? Journal of Developmental and Behavioral Pediatrics, 17, 22–26.

Geldhill, J., Ford, T., & Goodman, R. (2002). Does season of birth matter? The relationship between age within the school year (season of birth) and educational difficulties amongst a representative general population sample of children and adolescents (aged 5–15) in Great Britain. Research in Education, 68, 41–47.

Goodman, R., Gledhill, J., & Ford, T. (2003). Child psychiatric disorder and relative age within school year: cross-sectional survey of large population sample. British Medical Journal, 327, 472–475.

Grigorenko, E. L. (2001). Developmental dyslexia; an update on genes, brains and environments. Journal of Child Psychology and Psychiatry, 42, 91–125.

Hollingshead, A. (1957). Four factor index of social status. New Haven: Yale University Department of Sociology.

Jastak, S., & Wilkinson, G. S. (1984). Administration manual: the wide range achievement test-revised. Wilmington, DE: Jastak Associates Inc.

Job, R., Sartori, G., & Tressoldi, P. E. (1995). Battery for the evaluation of dyslexia and dysorthographia. Firenze, Italy: Organizzazioni Speciali.

Lawlor, D. A., Clark, H., Ronalds, G., & Leon, D. A. (2006). Season of birth and childhood intelligence: findings from the Aberdeen Children of the 1950s cohort study. The British Journal of Educational Psychology, 76(Pt 3), 481–499.

Levitan, R. D., Masellis, M., Lam, R. W., Kaplan, A. S., Davis, C., Tharmalingam, S., et al. (2006). A bith- season/DRD4 gene interaction predicts weight gain and obesity in women with seasonal affective disorder; a seasonal thrifty phenotype hypothesis. Neuropsychopharmacology, 31, 2498–2503.

Liederman, J., Kantrowitz, L., & Flannery, K. (2005). Male vulnerability to reading disability is not likely to be a myth: a call for new data. Journal of Learning Disabilities, 38, 473–477.

Lindgren, D. S., De Renzi, E., & Richman, L. C. (1985). Cross-national comparison of developmental dyslexia in Italy and the United States. Child Development, 56, 1404–1417.

Livingston, R., Adam, B. S., & Bracha, H. S. (1993). Season of birth and neurodevelopmental disorders: summer birth is associated with dyslexia. Journal of the American Academy of Child and Adolescent Psychiatry, 32, 612–616.

Lyon, G. R. (1996). Learning disabilities. The Future of Children, 6, 54–76. Shaywitz, S. E. (1998). Dyslexia. The New England Journal of Medicine, 338, 307–312. Shaywitz, S. E., & Shaywitz, B. A. (2005). Dyslexia (specific reading disability). Biological Psychiatry, 57,

1301–1309. Shaywitz, S. E., Shaywitz, B. A., Fletcher, J. M., & Escobar, M. D. (1990). Prevalence of reading disability

in boys and girls. Results of the Connecticut Longitudinal Study. Journal of the American Medical Association, 264, 998–1002.

Is season of birth related to developmental dyslexia? 181

St Sauver, J. L., Katusic, S. K., Barbaresi, W. J., Colligan, R. C., & Jacobsen, S. J. (2001). Boy/girl differences in risk for reading disability: potential clues? American Journal of Epidemiology, 154, 787– 794.

Tressoldi, P. E., Stella, G., & Faggella, M. (2001). The development of reading speed in Italians with dyslexia: a longitudinal study. Journal of Learning Disabilities, 34, 414–417.

Wechsler, D. (1986). Wechsler intelligence scale for children revised. Firenze: Organizzazioni Speciali. Williams, J., & O'Donovan, M. C. (2006). The genetics of developmental dyslexia. European Journal of

Human Genetics, 14, 681–689.

182 R. Donfrancesco et al.

Copyright of Annals of Dyslexia is the property of Springer Science & Business Media B.V. and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.