We show, for the first time, that maternal grandfather’s age is associated with risk of schizophrenia. Compared to the reference category (20 to 24 years), the offspring of women with paternal age of 40 years or more, have an increased risk of schizophrenia Unexpectedly, we find that paternal grandfathers’s age is not associated with risk of schizophrenia. In keeping with previous studies (Byrne et al., 2003
; Malaspina et al., 2001
; Sipos et al., 2004
), we confirm a statistically significant association between advanced paternal age and an increased offspring risk of schizophrenia.
The selective effect for maternal but not paternal grandpaternal age may provide clues to the biological mechanisms underpinning the link between advanced paternal age and risk of schizophrenia. One key difference between maternal versus paternal grandfathers is the segregation of the X chromosome. The paternal grandfather’s X chromosome is not inherited by his son, nor any of this son’s children (i.e. grandsons and granddaughters). In contrast, the maternal grandfather’s X chromosome is inherited by his daughter and to half of her sons, and half of her daughters. In many instances, X-linked transmission is associated with disease phenotypes emerging in males only (for X-linked recessive) or in a more prominent fashion in males compared to females (for X-linked dominant). While the incidence of schizophrenia has been reported to be higher in men compared to women (Aleman et al., 2003
; McGrath et al., 2008
) and there are a range of sex differences in the features of schizophrenia in women versus men (Leung and Chue, 2000
), when we examined the variables of interest in men and women separately, we did not detect any appreciable difference according to the sex of the proband. A study based on Danish mental health registries identified, in fathers older than 50 years, a higher risk of schizophrenia in female compared to male offspring (Byrne et al., 2003
). The authors also speculated that paternal age might specifically impact on X-chromosome linked genes. A study based on the Jerusalem birth cohort has also reported a substantially increased risk of schizophrenia in the sisters of women with schizophrenia who have older fathers – this pattern of affected female sibpairs suggests that paternal age may differentially impact on the X chromosome (Perrin et al., 2010
). A recent systematic review did not detect any sex difference in risk of schizophrenia in the offspring of older fathers (Miller et al., 2010
). Our findings related to the specificity of association with maternal versus paternal grandfather adds weight to the hypothesis that the X chromosome may be differentially affected by age-related mutagenesis in the male germ line. However, on first principles, age-related de novo
mutations should impact on autosomes as well as the sex chromosomes. Thus, the lack of association between paternal grandfathers age and risk of schizophrenia is hard to explain.
Recently, it has been demonstrated that the expression of genes on maternal versus paternal chromosomes operate in different regions of the brain during development and in adulthood. For example, genes on maternal chromosomes are preferentially transcribed in the developing brain, while genes from paternal chromosomes are preferentially transcribed in adult brain (Gregg et al., 2010a
). To further complicate the complex links between genotype and phenotype, there are also differences in maternal versus paternal genome transcription in male versus female offspring (Gregg et al., 2010b
). There is also evidence that candidate single nucleotide polymorphisms (SNPs) within known imprinted regions of the genome can be either protective or harmful for disease outcomes depending on the parent of origin of the variant (Kong et al., 2009
) (Kong et al., 2009
). This type of biological complexity cannot be unravelled with epidemiology (McGrath and Richards, 2009
), but recently developed animal models related to advanced paternal age may provide clues to underlying mechanisms (Foldi et al., 2010
; Smith et al., 2009
The main strengths of the present study were the large sample size, the nation-wide coverage of psychiatric inpatient care in Sweden, the standardized routines for ICD diagnostic reporting in Sweden (which optimizes the reliability of case ascertainment) (Ekholm et al., 2005
) and access to the Multigeneration Register. However, the study has several limitations. The registers are currently limited in respect to follow-up time for the three-generation study resulting in a much smaller sample with present grandparental age data compared to the sample with parental age data. Furthermore, our three-generation sample would over-represent early-onset patients and/or those with younger parents. With respect age of onset, there is some evidence to suggest that advanced paternal age is associated with an earlier age of onset of schizophrenia (Lee et al., 2011
; Rosenfield et al., 2010
), so this feature may lead to over-estimates of effect size in our sample. With respect to a possible over-representation of younger second-generation parents, some studies examining paternal age and risk of schizophrenia have identified a J-shaped curve, with a slight excess risk in the offspring of younger fathers as well as a larger effect for older fathers (Miller et al., 2010
). It is not clear how this feature would impact on our findings – it is plausible that different biological and/or psychosocial mechanisms underlie the J shaped curve. Regardless of this speculation, it will be important to see if the pattern of associations persists as the Swedish register can access more third-generation pedigrees in the years ahead. It will also be of interest to explore if other neurodevelopmental disorders associated with paternal age also show differential maternal versus paternal grandfather effects. In particular, the association between grandpaternal age and disorders such as autism would be better suited to the register-based studies, as more three-generation pedigrees could be identified with an early onset disorder.
In conclusion, as well as replicating earlier findings of a positive association between advanced paternal age and schizophrenia, we have found that the advanced age of the maternal but not paternal grandfather is associated with an increased risk of schizophrenia in the grandchild.