Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Feb 1.
Published in final edited form as: Nord J Psychiatry. 2022 May 30;77(2):179–187. doi: 10.1080/08039488.2022.2078998

School performance and genetic propensities for educational attainment and depression in the etiology of self-harm: a Danish population-based study

Holger J Sørensen 1,2,11,*, Sussie Antonsen 1,3, Michael E Benros 1,2, Annette Erlangsen 1,2,8,9,10,*, Clara Albiñana 1,3, Merete Nordentoft 1,2,11, Anders D Børglum 1,6,14, Ole Mors 1,7, Thomas Werge 1,11,12, Preben B Mortensen 1,3, David Hougaard 1,13, Roger T Webb 5,*, Esben Agerbo 1,3,4,*
PMCID: PMC9883111  NIHMSID: NIHMS1862234  PMID: 35635301

Abstract

Background:

Poor school performance is linked to higher risks of self-harm. The association might be explained through genetic liabilities for depression or educational attainment. We investigated the association between school performance and self-harm in a population-based sample, while assessing the potential influence of polygenic risk scores (PRSs) for depression (PRSMDD) and for educational attainment (PRSEDU).

Method:

We conducted a follow-up study of individuals born 1987–98 and followed from age 18 until 2016. The total sample consisted of a case group (23,779 diagnosed with mental disorders; schizophrenia, bipolar disorder, depression, autism, and attention deficit hyperactivity disorder (ADHD) and a randomly sampled comparison group (n=10,925). Genome-wide data were obtained from the Neonatal Screening Biobank and information on school performance, family psychiatric history, and socioeconomic status from national administrative registers.

Results:

Individuals in the top PRSMDD decile were at higher self-harm risk in the case group (IRR: 1.30; 95% CI 1.15–1.46), whereas individuals in the top PRSEDU decile were at lower self-harm risk (IRR: 0.63; 95% CI: 0.55–0.74). Poorer school performance was associated with higher self-harm risk in persons diagnosed with any mental disorder (IRR: 1.69; 95% CI: 1.44–1.99) and among the comparison group (IRR: 7.93; 95% CI: 4.47–15.18). Observed effects of PRSMDD and PRSEDU on self-harm risk were strongest for individuals with poor school performance.

Conclusion:

Associations between PRSMDD and self-harm risk and between PRSEDU and self-harm risk were found. Nevertheless, these polygenic scores seem currently of limited clinical utility for identifying individuals at high self-harm risk.

Keywords: School-performance, self-harm, mental disorders, polygenic risk scores, linkage data

INTRODUCTION

Self-harm incidence rates are highest among females during adolescence and in young adulthood.(Morthorst et al., 2016) Self-harm is linked to elevated risk of subsequently dying by suicide(Carroll et al., 2014) and therefore a major public health concern. The transmission of self-harm across generations is moderately strong and linked to both genetic and rearing effects.(Kendler et al., 2020) Mental disorders are associated with higher self-harm risk,(Erlangsen et al., 2020) and psychological, familial, sociocultural factors contribute to self-harm risk.(Hawton et al., 2012)

Several epidemiological studies have found that risks of self-harm vary according to educational gradient with an association between poorer school performance and self-harm,(Jablonska et al., 2009, Kosidou et al., 2014, Sörberg Wallin et al., 2018) although the association might largely be explained by sociodemographic confounders.(Björkenstam et al., 2011, Jablonska et al., 2009, Jablonska et al., 2012) However this association might be moderated by less modifiable genetic factors, and further analyses of these associations could help identify specific sub-types at elevated risk of self-harm.(Brunner et al., 2007)

Attention has been given to genetic components underlying self-harm in individuals with mental disorder.(Levey et al., 2019, Mullins et al., 2019, Ruderfer et al., 2020, Strawbridge et al., 2019) One of these studies found a SNP (single nucleotide polymorphism) heritability of 4.6%.(Erlangsen et al., 2020) It has been suggested that polygenic risk scores (PRS) might help identify people at risk of self-harm.(Laursen et al., 2017, Mullins et al., 2019) For instance, the PRS for major depressive disorder (MDD) (PRSMDD) has been linked to self-harm risk in individuals with depression, bipolar disorder and schizophrenia,(Mullins et al., 2019) whereas PRS for schizophrenia may not be associated with self-harm risk.(Laursen et al., 2017) Findings from a large cohort have provided new information regarding associations between genetic variants and risk of suicidal thoughts and behaviors.(Strawbridge et al., 2019) Furthermore, mode of genetic covariation has been positively correlated with polygenic scores for a number of mental disorders, while on the other hand, being negatively correlated with the polygenic risk score for educational attainment.(Taquet et al., 2021) However, it remained to be examined whether polygenic scores for educational attainment (PRSEDU) are associated with self-harm.(Selzam et al., 2017, Sørensen et al., 2018)

Using Danish nationwide registers, our aims were: (1) to investigate whether PRSMDD or PRS EDU are associated with self-harm, and (2) to examine whether PRSMDD or PRSEDU can explain the association between poor school performance and self-harm.

MATERIALS AND METHODS

Data sources

The study was conducted by implementing a nested design and using data from multiple Danish population-based registers, which were linked using the unique personal identification numbering system introduced in 1968.(Pedersen, 2011) Data were obtained from following registers: the Neonatal Screening Biobank, which contains information from dried blood spots taken 5–7 days after birth from virtually all infants born in Denmark after 1981,(Nørgaard-Pedersen and Hougaard, 2007) and the Danish Psychiatric Central Research Register, which contains data on all psychiatric inpatient admissions since 1969 and all outpatient treatment and emergency room visits since 1995.(Mors et al., 2011) Diagnoses were recorded according to the 8th Revision of the International Classification of Diseases (ICD-8) from 1977 to 1993, and ICD-10 from 1994.(World Health, 1993) ICD-9 was never implemented in Denmark.

Study cohorts

The case group consisted of singletons born between January 1, 1987 to December 31, 1998, who were alive and residing in Denmark at their 18th birthday and had two Danish-born parents. This cohort was derived from the iPSYCH2012 sample, consisting individuals diagnosed with schizophrenia (F20), bipolar disorder (F30, 31), mood disorders (F32–33), autism spectrum disorder (F84.0, F84.1, F84.5, F84.8) and attention-deficit hyperactivity disorder (ADHD) (F90.0) (supplemental online material, Table 1),(Pedersen et al., 2018) in addition to a comparison group of population-based controls.(Pedersen et al., 2018) Only mental disorders that occurred before the outcome (i.e. self-harm) were considered, but a mental disorder may occur both before and after the time of school performance. After control for genetic outliers and passing quality control,(Grove et al., 2019) PRS was generated for 34,184 individuals (supplemental online material, Table 2).

Exposure measurements

The PRS is a sum of autosomal risk alleles carried by an individual, where each term is weighted by the corresponding log-odds ratio from the specific separate discovery GWAS sample. (Wray et al., 2014) The PRS for educational attainment (PRSEDU) was GWAS-based with summary statistics (n=1,131,881 individuals).(Lee et al., 2018) The PRSEDU correlates positively with years of schooling, so that individuals with more years of schooling tend to have a higher score. Summary statistics were filtered at minor allele frequency (MAF) of 0.05 and info score at 0.9, when present, and we removed the broad major histocompatibility complex-region (MHC-region) (chromosome 6: 25–35MB). All ambiguous markers were removed to avoid potential strand conflicts. The summary statistics were restricted to SNPs known to be present in iPSYCH data at info score ≥ 0.01 throughout all 23 sample waves (batches). Clumping was done on the filtered summary statistics employing the whole genome association analysis tool set, PLINK. Details about iPSYCH genotyping, ancestry filtering and quality control (QC) has previously been described. PRSMDD were derived using the most recently published results from the Psychiatric Genomics Consortium (not including iPSYCH2012) as discovery GWAS datasets.(Howard et al., 2019) Both scores were standardized using arithmetic mean and standard deviation in the population controls only.

We obtained data on school performance from the Primary Education Register.(Jensen and Rasmussen, 2011) Primary education lasts for 10 years in Denmark and is comparable to elementary/middle school in North America and primary/lower-secondary school in the United Kingdom. Exams are mandatory and grades are based on agreement between the teacher and an external examiner and registered on a 7-point scale with 12 being the highest grade (individual grades: −3, 00, 02, 4, 7, 10, and 12). Mean grades, typically given at age 16, were reported as Grade Point Averages (GPAs) for the following 3 exams: 1) written mathematics, 2) oral mathematics and 3) Danish. Individuals who completed all 3 exams with a GPA of at least ‘02’ before July 1st the year they turned age 17 (corresponding to 99% of those who were tested in all 3 exams) were considered as having successful completed primary school. Information on non-completion of primary education was also obtained. The measured exposure, school performance, was categorized into these 6 groups: non-completers, and quintiles of GPA with the highest quintile being the reference category.

Outcome

Self-harm was defined using a coding algorithm applied to all hospital-treated episodes recorded in the National Patient Register(Lynge et al., 2011) and Psychiatric Central Research Register(Mors et al., 2011) (supplemental online material, Table 3). This approach has previously been applied.(Nordentoft et al., 2011) Information on self-harm events was available until December 31st, 2016.

Covariates

Parental psychiatric history was identified in the Psychiatric Central Research Register as any diagnosed mental disorder (ICD-8: 290–315; ICD-10: F00-F99). Data on parental socioeconomic position (SEP) for the year when cohort members’ reached age 17 years was extracted from the Integrated Database for Labour Market Research.(Petersson et al., 2011) ‘Lower SEP’ was defined as both parents fulfilling at least one of following three criteria: lowest income quintile, highest education being primary school, outside the workforce. ‘Higher SEP’ was assigned when both mother and father were employed and at least one of the following criteria was fulfilled: highest income quintile or having higher education. ‘Middle SEP’ was defined as all other possible combinations.

Statistical analysis

Individuals were followed from their 18th birthday until the first self-harm event, death, emigration or the study’s final observation date (December 31, 2016), whichever came first. Incidence rate ratios (IRRs) and their 95% confidence intervals (CIs) for first self-harm episode were estimated by log-linear Poisson regression using the SAS 9.4 GENMOD procedure. The basic multivariable model was adjusted for sex, age and calendar year and the final model for PRSEDU and PRSMDD, parental psychiatric history and parental SEP. Age, calendar year and parental psychiatric history were treated as time-dependent variables, and all other covariates as time-fixed.

The number of self-harm incidents during follow-up were truncated to 0, 1 or 2 or more events and analysed in a multinomial logistic regression model using the generalized logit function in the SAS 9.4 LOGISTIC procedure. To control for population stratification, all models that included PRSs were also adjusted for the first two ancestral principal components. All models were analysed separately for members of the case group, each mental disorder, and the comparison group. Population marginal or Least square (LS)(Searle et al., 1980) means adjusted for sex, year of birth and the first two principal components were calculated for PRSEDU and PRSMDD, separately for six categories of school performance (non-completers and GPA quintile) in persons with and without a self-harm incidents during follow-up.

RESULTS

The total sample consisted of 34,184 individuals (54.4% males) including a case group consisting of individuals diagnosed mental disorders (N=23,779 subdivided into 9,869 with depression, 8,116 with ADHD, 6,839 with autism and 1,334 with schizophrenia and 616 with bipolar disorder) and 10,925 individuals in the comparison group (Supplemental online material, Table 2).

Potential risk factors for self-harm

Female gender, a parental history of mental disorder, lower SEP and noncompletion of primary school were associated with elevated self-harm risk in both the case and comparison group (Table 1). Persons in the top PRSEDU decile had lower risks of self-harm compared to persons with lower PRSEDU scores. A substantially higher proportion in the case group (43.1%) did not complete the exam in primary school than in the comparison group (13.5%) (supplemental online material Table 4). We found a 6.4-fold higher incidence rate of self-harm among individuals with mental disorders when compared to the comparison group.

Table 1.

Number of events and incidence rates (IR, per 10,000 person-years) for self-harm for case and comparison group.

Case group Comparison group Total sample
n IR n IR n IR
Overall 2,389 175.9 175 27.6 2,506 127.7
Gender
Males 849 117.2 65 20.1 898 87.1
Females 1,540 242.8 110 35.5 1,608 172.8
Birth year
1987–1990 715 185.7 53 29.5 745 134.1
1991–1994 667 193.0 47 29.8 698 140.5
1995–1998 1,007 160.4 75 25.4 1,063 116.9
Primary school performance
Completers 1,449 178.7 117 21.5 1,530 114.4
Non completers 940 171.7 58 65.4 976 156.1
Grade point average, quintiles
1st (lowest) 372 204.6 30 33.6 393 147.7
2nd 377 205.7 37 37.5 402 145.1
3rd 220 166.1 16 18.5 234 108.1
4th 301 171.1 21 15.3 315 101.8
5th (highest) 179 130.3 13 9.8 186 69.4
History of parental mental illness
Yes 771 217.4 52 58.1 806 184.7
No 1,618 161.2 123 22.6 1,700 111.4
Parental socioeconomic position
Lower 603 212.8 40 46.8 630 173.3
Middle 1,381 173.1 106 27.0 1,452 123.9
Higher 214 116.7 16 12.7 224 73.4
Missing 191 204.0 13 44.1 200 164.5
PRS education
Highest decile 157 126.1 8 10.9 164 83.5
Decile 1–9 2,232 180.9 167 29.9 2,342 132.6
PRS major depression
Highest decile 302 201.2 18 36.5 315 161.0
Decile 1–9 2,087 172.7 157 26.9 2,191 124.0
Open in a new tab

Abbreviations: PRS, polygenic risk scores; IR, Incidence rate.

Polygenic risk scores

As indicated by the results in Table 2, we found an association between PRSMDD and self-harm risk in the case group. Cases with any mental disorder and with a PRSMDD in the top decile had significantly higher self-harm risk (IRR 1.30; 95% CI 1.15–1.46) compared to cases with any mental disorder and with lower PRSMDD scores. In the comparison group, persons in the top PRSMDD decile had a non-significantly elevated (IRR 1.36; 95% CI 0.81–2.15) self-harm risk compared to lower PRSMDD scores and an elevated risk (IRR 1.17; 95% CI 95% CI 1.03–1.31) was also found in the total sample. Results from Table 2 also indicate that persons in the top PRSEDU decile had significantly lower risk than lower PRSEDU scores in the case group (IRR 0.63; 95% CI 0.54–0.74) and, particularly, in the control group (IRR 0.36; 95% CI 0.16–0.69).

Table 2.

Incidence rate ratio for self-harm according to PRSMDD and PRSEDU for case and comparison group.

Case group
IRR (95% CI)		 Comparison group
IRR (95% CI)		 Total sample
IRR (95% CI)
PRS MDD
 Highest decile 1.30 (1.15–1.46) 1.36 (0.81–2.15) 1.17 (1.03–1.31)
 Decile 1–9 1.00 (reference) 1.00 (reference) 1.00 (reference)
PRS EDU
 Highest decile 0.63 (0.54–0.74) 0.36 (0.16–0.69) 0.70 (0.59–0.82)
 Decile 1–9 1.00 (reference) 1.00 (reference) 1.00 (reference)
Open in a new tab

Abbreviation: PRSEDU, polygenic risk scores for educational attainment; PRSMDD, polygenic risk scores for depression; IRR, Incidence Rate Ratio; CI, confidence interval.

Potential risk factors for self-harm in each of the 5 disorders

For each of the 5 mental disorders shown in Table 3, we found higher self-harm rates among those exposed to the risk factors (e.g. parental psychiatric history or low SEP) that had also shown association with self-harm risk in Table 1.

Table 3.

Number of events and incidence rates (IR, per 10,000 person-years) for self-harm with respect to specific disorders.

Schizophrenia Bipolar Depression Autism ADHD
n IR n IR n IR n IR n IR
Overall 316 373.1 119 271.1 1,556 236.9 215 66.3 657 163.1
Gender
Males 133 272.4 32 186.6 374 178.2 102 40.2 366 125.3
Females 183 510.3 87 325.2 1,182 264.5 113 159.2 291 262.9
Birth year
1987–1990 82 316.6 50 271.4 519 220.1 34 60.9 154 177.0
1991–1994 95 370.9 37 268.8 451 240.6 48 69.9 167 186.8
1995–1998 139 418.9 32 273.1 586 250.9 133 66.5 336 148.4
Primary school performance
Completers 183 318.8 79 239.7 1,069 217.1 94 73.0 260 147.6
Non-completion 133 487.4 40 365.6 487 296.3 121 61.8 397 175.2
Grade point average, quintiles
1st (lowest) 37 282.0 18 370.9 243 262.6 18 75.4 105 159.1
2nd 48 326.0 14 215.1 277 261.7 24 89.1 65 136.3
3rd 34 367.9 13 228.2 164 206.2 15 63.5 43 149.8
4th 45 375.5 19 245.3 245 204.8 20 71.0 28 120.6
5th (highest) 19 227.8 15 184.0 140 147.5 17 65.1 19 179.2
Parental mental illness mmental illness
Yes 118 480.0 39 281.5 471 275.7 58 76.2 234 203.3
No 198 329.4 80 266.2 1,085 223.2 157 63.2 423 147.0
Parental socioeconomic position
Lower 74 419.1 24 269.1 355 274.3 36 63.6 209 199.5
Middle 181 379.9 67 287.1 922 240.3 139 70.5 359 152.4
Higher 35 294.8 20 227.3 153 159.1 28 53.6 39 113.9
Missing 26 345.8 8 281.4 126 264.9 12 65.4 50 176.7
PRS education
Highest decile 24 304.5 13 259.4 113 184.1 19 44.3 29 139.9
Decile 1–9 292 380.1 106 272.6 1,443 242.3 196 69.6 628 164.4
PRS major depression
Highest decile 52 580.7 14 240.4 196 250.7 38 114.6 83 191.4
Decile 1–9 264 348.6 105 275.7 1,360 235.0 177 60.8 574 159.7
Open in a new tab

Abbreviations: ADHD, attention-deficit hyperactivity disorder; IR, Incidence rate.

Polygenic risk scores, school performance and self-harm

Associations between school performance and self-harm were found for the case group and comparison group as well as in the total sample (Table 4). The IRRs were highest among those with poor school performance and robust estimates were obtained in the fully adjusted analyses. In the total sample, people who did not complete primary school had a 2.89-fold higher rate of self-harm (95% CI 2.39–3.28) than those with the best school performance. Adjusting for PRSMDD yielded a comparable association (IRR 2.74; 95% CI 2.35–3.22), whereas adjustment for PRSEDU attenuated the association by 9.0 % (IRR 2.54; 95% CI 2.16–2.99). In the case group, those who did not complete primary school also had a higher self-harm rate (basic model: IRR: 1.69; 95% CI: 1.44–1.99; fully adjusted model: IRR: 1.44; 95% CI: 1.22–1.72) when compared to those who completed primary school. Adjustment for PRSMDD yielded a similar estimate (1.68; 95% CI 1.43–1.98) while PRSEDU attenuated the estimate by 8.0 % (to 1.55; 95% CI 1.32–1.84). Among those in the case group who finished primary school, the lowest quintile of GPA was linked to the highest rate of self-harm (fully adjusted IRR: 1.50; 95% CI 1.24–1.81) when compared to the highest GPA quintile. In the comparison group, non-completers had a markedly higher self-ham rate (IRR 7.38; 95% CI 4.31–13.50) in the basic adjustment model than those with the highest GPA quintile. Adjusting for PRSMDD, yielded a similar association (IRR 7.78; 95% CI 4.39–14.90) whereas adjustment for PRSEDU attenuated the association by 14.8% (IRR 6.29; 95% CI 3.48–12.20). In the fully adjusted model, non-completers had a higher rate of self-harm (IRR 4.63; 95% CI 2.48–9.21) compared to those who completed their primary school education with a GPA in the highest quintile. Individuals with a GPA in the lowest and next-lowest quintile also had a higher self-harm rate (IRR 2.56; 95% CI: 1.31–5.26 and 3.02; 95% CI 1.61–6.05, respectively).

Table 4:

Incidence rate ratios for self-harm according to school performance at primary school for the total sample, comparison group, and case group.

Grade point average (GPA) quintile Basic model
(95% CI)		 IRR adjusted for PRSEDU
(95% CI)		 IRR adjusted for PRSMDD
(95% CI)		 IRR adjusted for parental psychiatric history and SEP
(95% CI)		 IRR fully adjusted
(95% CI)
Total sample
Non-completers 2.89 (2.39–3.28) 2.54 (2.16–2.99) 2.74 (2.35–3.22) 2.36 (2.00–2.81) 2.22 (1.87–2.64)
1st (lowest) 2.35 (1.98–2.81) 2.14 (1.79–2.56) 2.34 (1.97–2.80) 2.02 (1.68–2.44) 1.91 (1.59–2.31)
2nd 2.25 (1.90–2.69) 2.08 (1.75–2.49) 2.23 (1.88–2.66) 2.02 (1.68–2.43) 1.92 (1.60–2.31)
3rd 1.66 (1.37–2.01) 1.57 (1.29–1.90) 1.64 (1.35–1.99) 1.56 (1.28–1.91) 1.50 (1.23–1.84)
4th 1.48 (1.24–1.78) 1.44 (1.20–1.72) 1.48 (1.23–1.77) 1.40 (1.16–1.70) 1.37 (1.14–1.66)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Case group
Non-completers 1.69 (1.44–1.99) 1.55 (1.32–1.84) 1.68 (1.43–1.98) 1.52 (1.28–1.81) 1.44 (1.22–1.72)
1st (lowest) 1.74 (1.46–2.09) 1.60 (1.34–1.92) 1.75 (1.46–2.09) 1.57 (1.30–1.90) 1.50 (1.24–1.81)
2nd 1.72 (1.44–2.06) 1.60 (1.34–1.92) 1.72 (1.44–2.05) 1.60 (1.33–1.94) 1.53 (1.27–1.86)
3rd 1.37 (1.13–1.68) 1.31 (1.07–1.60) 1.37 (1.12–1.67) 1.33 (1.09–1.64) 1.29 (1.05–1.59)
4th 1.34 (1.11–1.61) 1.30 (1.08–1.57) 1.33 (1.11–1.61) 1.28 (1.06–1.55) 1.25 (1.04–1.52)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Comparison group
Non-completers 7.93 (4.47–15.18) 6.29 (3.48–12.20) 7.78 (4.39–14.90) 5.51 (3.00–10.86) 4.63 (2.48–9.21)
1st (lowest) 3.96 (2.11–7.88) 3.14 (1.65–6.33) 3.91 (2.08–7.77) 3.04 (1.57–6.18) 2.56 (1.31–5.26)
2nd 4.24 (2.31–8.30) 3.52 (1.90–6.95) 4.21 (2.29–8.24) 3.43 (1.83–6.83) 3.02 (1.61–6.05)
3rd 2.03 (0.98–4.30) 1.76 (0.84–3.75) 2.01 (0.97–4.26) 1.71 (0.81–3.69) 1.54 (0.72–3.33)
4th 1.61 (0.82–3.31) 1.50 (0.76–3.08) 1.62 (0.82–3.33) 1.46 (0.73–3.01) 1.40 (0.70–2.89)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Open in a new tab

Abbreviation: PRSEDU, polygenic risk scores for educational attainment; PRSMDD, polygenic risk scores for depression; IRR, Incidence Rate Ratio; CI, confidence interval; SEP, parental socioeconomic position.

Polygenic risk scores, school performance and self-harm in specific mental disorders

Non-completion of primary school was associated with higher risk of self-harm in schizophrenia (IRR: 2.35; 95% CI: 1.45–4.05), bipolar disorder (IRR: 2.25; 95% CI: 1.17–4.52) and depression (IRR: 1.95; 95% CI: 1.59–2.40) when compared to those in the highest GPA quintile (Table 5). Among those who finished primary school, the lowest quintile of GPA was linked to significantly a higher rate of self-harm for bipolar disorder (IRR 2.50; 95% CI 1.15–2.50) and depression (IRR 1.70; 95% CI 1.36–2.13) in the fully adjusted analyses. This suggests that neither PRSMDD, PRSEDU scores, nor parental family history of mental illness nor low SEP explained the association between poor school performance and self-harm risk in bipolar disorder or depression.

Table 5.

Incidence rate ratios for self-harm according to school performance at primary school for specific mental disorders

Grade point average (GPA) quintile IRR basic model
(95% CI)		 IRR adjusted for PRSEDU
(95% CI)		 IRR adjusted for PRSMDD
(95% CI)		 IRR adjusted for parental psychiatric history and SEP
(95% CI)		 Fully adjusted IRR
(95% CI)
Schizophrenia
Non-completion 2.37 (1.50–3.97) 2.33 (1.47–3.93) 2.41 (1.53–4.04) 2.36 (1.46–4.05) 2.35 (1.45–4.05)
1st (lowest) 1.39 (0.81–2.47) 1.38 (0.80–2.47) 1.44 (0.83–2.56) 1.21 (0.68–2.21) 1.22 (0.68–2.25)
2nd 1.59 (0.95–2.77) 1.56 (0.93–2.74) 1.62 (0.96–2.83) 1.57 (0.92–2.80) 1.58 (0.92–2.83)
3rd 1.62 (0.94–2.90) 1.65 (0.95–2.96) 1.67 (0.96–2.98) 1.62 (0.92–2.94) 1.65 (0.94–3.00)
4th 1.73 (1.03–3.04) 1.77 (1.05–3.10) 1.77 (1.05–3.10) 1.63 (0.95–2.92) 1.67 (0.97–2.99)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Bipolar disorder
Non-completion 2.19 (1.23–4.11) 2.03 (1.10–3.92) 2.12 (1.19–3.98) 2.35 (1.25–4.62) 2.25 (1.17–4.52)
1st (lowest) 2.12 (1.07–4.28) 1.99 (0.97–4.16) 2.22 (1.11–4.48) 2.44 (1.16–5.22) 2.50 (1.15–5.51)
2nd 1.26 (0.60–2.63) 1.20 (0.56–2.56) 1.29 (0.61–2.70) 1.40 (0.64–3.02) 1.38 (0.62–3.04)
3rd 1.31 (0.61–2.75) 1.26 (0.58–2.69) 1.30 (0.61–2.74) 1.45 (0.67–3.15) 1.45 (0.65–3.18)
4th 1.35 (0.69–2.70) 1.30 (0.65–2.62) 1.36 (0.69–2.72) 1.41 (0.70–2.89) 1.40 (0.69–2.90)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Depression
Non-completion 2.14 (1.78–2.59) 2.06 (1.70–2.50) 2.13 (1.77–2.58) 1.99 (1.63–2.45) 1.95 (1.59–2.40)
1st (lowest) 1.85 (1.50–2.28) 1.78 (1.44–2.20) 1.85 (1.50–2.28) 1.73 (1.38–2.16) 1.70 (1.36–2.13)
2nd 1.84 (1.51–2.27) 1.79 (1.46–2.20) 1.84 (1.50–2.26) 1.75 (1.41–2.17) 1.72 (1.39–2.14)
3rd 1.43 (1.15–1.80) 1.41 (1.12–1.77) 1.43 (1.14–1.80) 1.41 (1.12–1.79) 1.40 (1.11–1.77)
4th 1.42 (1.15–1.75) 1.40 (1.14–1.73) 1.41 (1.15–1.74) 1.36 (1.10–1.69) 1.35 (1.09–1.68)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Autism
Non-completion 1.17 (0.72–2.02) 1.10 (0.68–1.92) 1.19 (0.73–2.05) 1.24 (0.74–2.23) 1.17 (0.70–2.12)
1st (lowest) 1.30 (0.66–2.54) 1.23 (0.63–2.43) 1.33 (0.68–2.60) 1.21 (0.58–2.50) 1.15 (0.55–2.40)
2nd 1.84 (0.99–3.50) 1.76 (0.95–3.36) 1.87 (1.01–3.55) 1.95 (1.02–3.83) 1.86 (0.97–3.66)
3rd 1.24 (0.61–2.49) 1.18 (0.58–2.39) 1.27 (0.62–2.55) 1.27 (0.61–2.65) 1.22 (0.58–2.55)
4th 1.19 (0.62–2.30) 1.16 (0.61–2.25) 1.20 (0.63–2.32) 1.34 (0.69–2.67) 1.30 (0.67–2.59)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
ADHD
Non-completers 1.14 (0.74–1.87) 1.08 (0.69–1.77) 1.14 (0.74–1.87) 1.07 (0.68–1.83) 1.04 (0.65–1.78)
1st (lowest) 0.95 (0.60–1.60) 0.90 (0.56–1.52) 0.96 (0.60–1.61) 0.90 (0.55–1.57) 0.88 (0.54–1.53)
2nd 0.81 (0.50–1.39) 0.78 (0.48–1.34) 0.81 (0.50–1.39) 0.78 (0.46–1.38) 0.76 (0.45–1.35)
3rd 0.89 (0.52–1.56) 0.86 (0.51–1.51) 0.89 (0.52–1.55) 0.84 (0.48–1.52) 0.82 (0.47–1.49)
4th 0.69 (0.39–1.25) 0.68 (0.38–1.23) 0.69 (0.39–1.25) 0.64 (0.35–1.20) 0.63 (0.34–1.19)
5th (highest) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref) 1.00 (ref)
Open in a new tab

Abbreviation: ADHD, attention-deficit hyperactivity disorder; PRSEDU, polygenic risk scores for educational attainment; PRSMDD, polygenic risk scores for depression; IRR, Incidence Rate Ratio; CI, confidence interval; SEP, parental socioeconomic position.

Further sub-group analysis by sex revealed that both male and female non-completers of primary school had a more than two-fold higher self-harm rates in the fully adjusted models (males: IRR 2.86; 95% CI 1.91, 4.52); females: IRR 2.26; 95% CI 1.87, 2.75).

DISCUSSION

First, we found that a PRSMDD in the top decile predicted elevated rates of self-harm among individuals diagnosed with any mental disorder. Second, we present novel findings showing that a PRSEDU in the top decile was associated with a lower self-harm rate when compared to lower deciles. This association was particularly strongly in the comparison group. A lower self-harm rate for individuals with a PRSEDU in the top decile was also found among individuals diagnosed with a mental disorder. Third, we found that poor school performance was associated with an increased rate of self-harm in the comparison and case group when compared to those with the highest GPA. Non-completion of primary school or finishing primary school with low school marks were both linked to higher rates of self-harm rate for people with bipolar disorder and depression measured relative to those with the highest GPA scores. The elevated rates of self-harm among individuals with poor school performance was found in both sexes and remained significant after adjustment for a wide range of potential confounders; indicating that the association between poor school performance and self-harm is not explained through genetic liabilities for depression or educational attainment.

Our results with respect to PRSMDD and self-harm corroborate the results of a recent GWAS and could support an interpretation where individuals with self-harm and mental disorder carry a greater liability for depression than individuals with mental disorder but without self-harm.(Mullins et al., 2019) It has been suggested that the genetic etiology of self-harm could be, at least partially, shared with major depression.(Mullins et al., 2019) To further explore self-harm as a genetically influenced phenotype among individuals with mental disorder would have required more powerful PRS and future studies adding genetic propensities for e.g. schizophrenia or ADHD(Demontis et al., 2019) could explore this in more detail, for instance using multi-traits prediction.(Grove et al., 2019)

We found genetic predisposition to educational attainment to correlate with lower rates of self-harm in both individuals with mental disorders and population-based comparisons. However, it is unclear whether the observed association is causal. Nevertheless, causal associations can be inferred from two sets of GWAS summary data from independent studies with large sample sizes by using the so-called GSMR method.(O’Connor and Price, 2018, Zhu et al., 2018) We did not find support that PRSEDU had a larger impact as a confounder on the association between school performance and self-harm than the combination of SEP and parental psychiatric history. This is in line with previous findings for major depression where an association between lower eeducational attainment and major depression was seemingly not related to measurable pleiotropic genetic effects. This suggests that environmental factors, for example socioeconomic status, might play a role.(Peyrot et al., 2015) Also, a study of a brain vulnerability network in children and adolescents has found an endophenotype funneling genetic risk for various psychiatric illnesses to be negatively correlated with PRS for educational attainment.(Taquet et al., 2021) PRS for educational attainment influence both schooling and important psycho-social characteristics(Belsky et al., 2016) and people who carry more alleles associated with educational attainment are found to have lower risks for meeting clinical criteria for alcohol, nicotine and cannabis use disorders.(Salvatore et al., 2020) Recent evidence supports a causal role of cannabis, alcohol and smoking on self-harm.(Orri et al., 2021)

Strengths of this study included the large, complete, and population-based sample, the use of prospectively collected data from accurately interlinked national registers on exposures, covariates and outcomes. Healthcare is accessible free of charge to all Danish residents, which largely eliminates selection bias. This design enabled us to estimate unbiased incidence rates and strengths of associations between genetic markers on risk of mental disorders without any potential for reverse causality between the PRSs and self-harm risk. We are not aware of previous studies of the association between school performance and self-harm risk based in similarly large data samples.

Limitations of this study included examination only of mental disorders that were diagnosed and treated in secondary care. Individuals with undiagnosed mental disorders or those treated by a general practitioner and without contact to secondary mental health services were not recorded. Case identification for certain conditions, such as depression, could therefore be biased towards more severe cases. For the association between school performance and self-harm, reverse causality cannot be excluded as preceding suicidal ideation or unrecorded self-harm episodes could have led to poor school performances. Likewise, un-recorded self-harm episodes may have preceded the first recorded psychiatric occurrence. This is, however, unlikely to explain the robust associations observed between non-completion of primary school and self-harm. Although PRSEDU and PRSMDD were derived from large samples, they may not be sufficiently powerful to capture the full genetic influence across the whole genome, which could lead to residual confounding.(Martin et al., 2019, Visscher et al., 2017) It is well known that the polygenic score is a blunt instrument.(Agerbo et al., 2021) Although polygenic risk scores are not yet ready for clinical use,(Torkamani et al., 2018) they may soon become clinically useful in psychiatry.(Murray et al., 2021, Wray et al., 2021) Lastly, self-harm is under-recorded in Denmark but this would likely render estimates in a conservative direction.(Morthorst et al., 2016)

In summary, the study provides further evidence that genetic factors influence risk of self-harm. Self-harm was positively associated with a high PRS depression score among people with mental disorder but negatively associated with a high PRS for educational attainment in people with mental disorder and in population-based comparisons. While little evidence for the clinical utility of the polygenic scores was provided, our study suggested that individuals with mental disorders and those who do not complete school are at clinically significant increased risk of committing self-harm.

Supplementary Material

Supp 1

FUNDING DETAILS

Genotyping of the iPSYCH2012 samples was supported by grant numbers R102-A9118 and R155-2014-1724 from the Lundbeck Foundation, the Stanley Foundation, the Simons Foundation (SFARI 311789), and the National Institutes of Mental Health (NIMH 5U01MH094432-02). The Danish National Biobank resource is supported by the Novo Nordisk Foundation. The authors gratefully acknowledge the Psychiatric Genomics Consortium (PGC), the Social Science Genetic Association Consortium (SSGAC) and the research participants and employees of 23andMe, Inc. for providing the summary statistics used to general the polygenic risk scores.

Biographies

Associate professor Holger J. Sørensen is chief physician at Mental Health Centre Copenhagen. He has a H-index of 32. His scientific focus is epidemiology, alcohol use disorder, high-risk studies and premorbid cognitive function in relation to mental disorder.

Sussie Antonsen, MSc, is a research assistant at the National Centre for Register-based Research, Aarhus University, Denmark. She has more than 14 years of experience in working with data management and statistical analyses of the Danish national health registers.

Prof. Michael E. Benros, MD, PhD, is chief physician and professor at Mental Health Centre Copenhagen and the University of Copenhagen. He is among the Clarivate highly cited researchers and board member of several international organizations. His scientific focus is epidemiology, immunopsychiatry, genetics, and precision psychiatry.

Annette Erlangsen PhD is Head of Program at Danish Research Institute for Suicide Prevention, Mental Health Centre Copenhagen. She is affiliated to Johns Hopkins School of Public Health, USA and Australian National University. Her expertise is population-based and intervention studies with a focus on suicide prevention.

Clara Albiñana, MSc, BSc is a PhD student in Statistical Genetics at the graduate school of Health at Aarhus University, Denmark. Clara is supervised by Bjarni J. Vilhjalmsson and has over 10 peer-review publications in the use of polygenic risk scores in epidemiology and genetics of mental disorders.

Merete Nordentoft is a clinical psychiatrist and Professor of Psychiatry, University of Copenhagen. She is an expert in epidemiology, suicidal behaviour, psychopathology and early intervention in psychosis. She has worked with suicide prevention at a national level since 1997, and she is involved in unravelling excess mortality in mental illness.

Anders Børglum is Professor of Medical Genetics and Chair of Personalized Medicine Research at Department of Biomedicine, Aarhus University. His research focuses on identifying genes that confer risk or resilience to psychiatric disorders, functional characterization of the identified genes and translating the genetic insights to advance precision medicine in psychiatry.

Ole Mors is a Professor in the Psychosis Research Unit, Department of Clinical Medicine, University of Aarhus, Denmark. His research focuses on identifying genetic and environmental risk factors for mental disorders.

Our unit consists of psychiatrists, nurses, psychologists, molecular biologists and statisticians, totalling twenty researchers, including a professor, a consultant, four associate professors, four postdocs, seven PhD students and three research assistants.

Thomas Werge is the director of Institute of Biological Psychiatry (Copenhagen University Hospital), and professor of psychiatry (University of Copenhagen). He is an expert in psychiatric and complex traits genetic and epidemiology, a founder of numerous international disease genomics initiatives and leading translational efforts in clinical psychiatry.

Preben Bo Mortensen is a Professor and Head of the National Center for Register-Based Research at Aarhus University, Denmark. He is an expert in psychiatric epidemiology, suicidology, psychiatric genetics, and related socioeconomic and demographic components. He has conducted research in psychiatric epidemiology and psychiatric genetics for more than 35 years.

David M. Hougaard, MD, DMSC is Head of Danish Center for Neonatal Screening and Director Department for Congenital Disorders at Statens Serum Institute, Denmark. He has scientifically collaborated with several institutions in USA and Europe with focus on neonatal screening and cell biology as well as genetic, epigenetic and biomarker connections to metabolic, endocrine, psychiatric and cancer disorders.

Roger Webb joined the University of Manchester during 1998, where he received his PhD in Epidemiology in 2002. He is currently Professor of Mental Health Epidemiology. He conducts population-based studies of adverse outcomes in people diagnosed with mental illnesses, with a particular interest in investigating the determinants of non-fatal self-harm and suicide.

Professor Esben Agerbo, DrMedSc, MSc, BSc is employed at the National Centre for Register-based Research, Aarhus University, Denmark. Dr Agerbo (H-index 62) has published more than 270 peer-reviewed publications that cited over 20,000 times. Dr Agerbo’s scientific focus is the epidemiology and genetics of mental disorders and suicidal behaviour.

Footnotes

DISCLOSURE STATEMENT

The authors declare no financial interests or potential conflicts of interest.

DATA AVAILABILITY STATEMENT

Due to safety regulations, data are not available to share.

REFERENCES

  1. AGERBO E, TRABJERG BB, BORGLUM AD, SCHORK AJ, VILHJALMSSON BJ, PEDERSEN CB, HAKULINEN C, ALBINANA C, HOUGAARD DM, GROVE J, MCGRATH JJ, BYBJERG-GRAUHOLM J, MORS O, PLANA-RIPOLL O, WERGE T, WRAY NR, MORTENSEN PB & MUSLINER KL 2021. Risk of Early-Onset Depression Associated With Polygenic Liability, Parental Psychiatric History, and Socioeconomic Status. JAMA Psychiatry. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. BELSKY DW, MOFFITT TE, CORCORAN DL, DOMINGUE B, HARRINGTON H, HOGAN S, HOUTS R, RAMRAKHA S, SUGDEN K, WILLIAMS BS, POULTON R & CASPI A 2016. The Genetics of Success:How Single-Nucleotide Polymorphisms Associated With Educational Attainment Relate to Life-Course Development. Psychological Science, 27, 957–972. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. BJÖRKENSTAM C, WEITOFT GR, HJERN A, NORDSTRÖM P, HALLQVIST J & LJUNG R 2011. School grades, parental education and suicide—a national register-based cohort study. Journal of Epidemiology and Community Health, 65, 993–998. [DOI] [PubMed] [Google Scholar]
  4. BRUNNER R, PARZER P, HAFFNER J, STEEN R, ROOS J, KLETT M & RESCH F 2007. Prevalence and Psychological Correlates of Occasional and Repetitive Deliberate Self-harm in Adolescents. Archives of Pediatrics & Adolescent Medicine, 161, 641–649. [DOI] [PubMed] [Google Scholar]
  5. CARROLL R, METCALFE C & GUNNELL D 2014. Hospital presenting self-harm and risk of fatal and non-fatal repetition: systematic review and meta-analysis. PloS one, 9, e89944–e89944. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. DEMONTIS D, WALTERS RK, MARTIN J, MATTHEISEN M, ALS TD, AGERBO E, BALDURSSON G, BELLIVEAU R, BYBJERG-GRAUHOLM J, BÆKVAD-HANSEN M, CERRATO F, CHAMBERT K, CHURCHHOUSE C, DUMONT A, ERIKSSON N, GANDAL M, GOLDSTEIN JI, GRASBY KL, GROVE J, GUDMUNDSSON OO, HANSEN CS, HAUBERG ME, HOLLEGAARD MV, HOWRIGAN DP, HUANG H, MALLER JB, MARTIN AR, MARTIN NG, MORAN J, PALLESEN J, PALMER DS, PEDERSEN CB, PEDERSEN MG, POTERBA T, POULSEN JB, RIPKE S, ROBINSON EB, SATTERSTROM FK, STEFANSSON H, STEVENS C, TURLEY P, WALTERS GB, WON H, WRIGHT MJ, ANDREASSEN OA, ASHERSON P, BURTON CL, BOOMSMA DI, CORMAND B, DALSGAARD S, FRANKE B, GELERNTER J, GESCHWIND D, HAKONARSON H, HAAVIK J, KRANZLER HR, KUNTSI J, LANGLEY K, LESCH KP, MIDDELDORP C, REIF A, ROHDE LA, ROUSSOS P, SCHACHAR R, SKLAR P, SONUGA-BARKE EJS, SULLIVAN PF, THAPAR A, TUNG JY, WALDMAN ID, MEDLAND SE, STEFANSSON K, NORDENTOFT M, HOUGAARD DM, WERGE T, MORS O, MORTENSEN PB, DALY MJ, FARAONE SV, BØRGLUM AD & NEALE BM 2019. Discovery of the first genome-wide significant risk loci for attention deficit/hyperactivity disorder. Nat Genet, 51, 63–75. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. ERLANGSEN A, APPADURAI V, WANG Y, TURECKI G, MORS O, WERGE T, MORTENSEN PB, STARNAWSKA A, BØRGLUM AD, SCHORK A, NUDEL R, BÆKVAD-HANSEN M, BYBJERG-GRAUHOLM J, HOUGAARD DM, THOMPSON WK, NORDENTOFT M & AGERBO E 2020. Genetics of suicide attempts in individuals with and without mental disorders: a population-based genome-wide association study. Molecular Psychiatry, 25, 2410–2421. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. GROVE J, RIPKE S, ALS TD, MATTHEISEN M, WALTERS RK, WON H, PALLESEN J, AGERBO E, ANDREASSEN OA, ANNEY R, AWASHTI S, BELLIVEAU R, BETTELLA F, BUXBAUM JD, BYBJERG-GRAUHOLM J, BÆKVAD-HANSEN M, CERRATO F, CHAMBERT K, CHRISTENSEN JH, CHURCHHOUSE C, DELLENVALL K, DEMONTIS D, DE RUBEIS S, DEVLIN B, DJUROVIC S, DUMONT AL, GOLDSTEIN JI, HANSEN CS, HAUBERG ME, HOLLEGAARD MV, HOPE S, HOWRIGAN DP, HUANG H, HULTMAN CM, KLEI L, MALLER J, MARTIN J, MARTIN AR, MORAN JL, NYEGAARD M, NÆRLAND T, PALMER DS, PALOTIE A, PEDERSEN CB, PEDERSEN MG, DPOTERBA T, POULSEN JB, POURCAIN BS, QVIST P, REHNSTRÖM K, REICHENBERG A, REICHERT J, ROBINSON EB, ROEDER K, ROUSSOS P, SAEMUNDSEN E, SANDIN S, SATTERSTROM FK, DAVEY SMITH G, STEFANSSON H, STEINBERG S, STEVENS CR, SULLIVAN PF, TURLEY P, WALTERS GB, XU X, STEFANSSON K, GESCHWIND DH, NORDENTOFT M, HOUGAARD DM, WERGE T, MORS O, MORTENSEN PB, NEALE BM, DALY MJ & BØRGLUM AD 2019. Identification of common genetic risk variants for autism spectrum disorder. Nat Genet, 51, 431–444. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. HAWTON K, SAUNDERS KE & O’CONNOR RC 2012. Self-harm and suicide in adolescents. Lancet, 379, 2373–82. [DOI] [PubMed] [Google Scholar]
  10. HOWARD DM, ADAMS MJ, CLARKE TK, HAFFERTY JD, GIBSON J, SHIRALI M, COLEMAN JRI, HAGENAARS SP, WARD J, WIGMORE EM, ALLOZA C, SHEN X, BARBU MC, XU EY, WHALLEY HC, MARIONI RE, PORTEOUS DJ, DAVIES G, DEARY IJ, HEMANI G, BERGER K, TEISMANN H, RAWAL R, AROLT V, BAUNE BT, DANNLOWSKI U, DOMSCHKE K, TIAN C, HINDS DA, TRZASKOWSKI M, BYRNE EM, RIPKE S, SMITH DJ, SULLIVAN PF, WRAY NR, BREEN G, LEWIS CM & MCINTOSH AM 2019. Genome-wide meta-analysis of depression identifies 102 independent variants and highlights the importance of the prefrontal brain regions. Nat Neurosci, 22, 343–352. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. JABLONSKA B, LINDBERG L, LINDBLAD F, RASMUSSEN F, OSTBERG V & HJERN A 2009. School performance and hospital admissions due to self-inflicted injury: a Swedish national cohort study. Int J Epidemiol, 38, 1334–41. [DOI] [PubMed] [Google Scholar]
  12. JABLONSKA B, LINDBLAD F, ÖSTBERG V, LINDBERG L, RASMUSSEN F & HJERN A 2012. A national cohort study of parental socioeconomic status and non-fatal suicidal behaviour-the mediating role of school performance. BMC Public Health, 12, 17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. JENSEN VM & RASMUSSEN AW 2011. Danish education registers. Scandinavian Journal of Public Health, 39, 91–94. [DOI] [PubMed] [Google Scholar]
  14. KENDLER KS, OHLSSON H, SUNDQUIST J, SUNDQUIST K & EDWARDS AC 2020. The Sources of Parent-Child Transmission of Risk for Suicide Attempt and Deaths by Suicide in Swedish National Samples. Am J Psychiatry, 177, 928–935. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. KOSIDOU K, DALMAN C, FREDLUND P, LEE BK, GALANTI R, ISACSSON G & MAGNUSSON C 2014. School performance and the risk of suicide attempts in young adults: a longitudinal population-based study. Psychol Med, 44, 1235–43. [DOI] [PubMed] [Google Scholar]
  16. LAURSEN TM, TRABJERG BB, MORS O, BØRGLUM AD, HOUGAARD DM, MATTHEISEN M, MEIER SM, BYRNE EM, MORTENSEN PB, MUNK-OLSEN T & AGERBO E 2017. Association of the polygenic risk score for schizophrenia with mortality and suicidal behavior - A Danish population-based study. Schizophr Res, 184, 122–127. [DOI] [PubMed] [Google Scholar]
  17. LEE JJ, WEDOW R, OKBAY A, KONG E, MAGHZIAN O, ZACHER M, NGUYEN-VIET TA, BOWERS P, SIDORENKO J, KARLSSON LINNÉR R, FONTANA MA, KUNDU T, LEE C, LI H, LI R, ROYER R, TIMSHEL PN, WALTERS RK, WILLOUGHBY EA, YENGO L, AGEE M, ALIPANAHI B, AUTON A, BELL RK, BRYC K, ELSON SL, FONTANILLAS P, HINDS DA, MCCREIGHT JC, HUBER KE, LITTERMAN NK, MCINTYRE MH, MOUNTAIN JL, NOBLIN ES, NORTHOVER CAM, PITTS SJ, SATHIRAPONGSASUTI JF, SAZONOVA OV, SHELTON JF, SHRINGARPURE S, TIAN C, VACIC V, WILSON CH, OKBAY A, BEAUCHAMP JP, FONTANA MA, LEE JJ, PERS TH, RIETVELD CA, TURLEY P, CHEN G-B, EMILSSON V, MEDDENS SFW, OSKARSSON S, PICKRELL JK, THOM K, TIMSHEL P, VLAMING RD, ABDELLAOUI A, AHLUWALIA TS, BACELIS J, BAUMBACH C, BJORNSDOTTIR G, BRANDSMA JH, CONCAS MP, DERRINGER J, FURLOTTE NA, GALESLOOT TE, GIROTTO G, GUPTA R, HALL LM, HARRIS SE, HOFER E, HORIKOSHI M, HUFFMAN JE, KAASIK K, KALAFATI IP, KARLSSON R, KONG A, LAHTI J, VAN DER LEE SJ, LEEUW CD, LIND PA, LINDGREN K-O, LIU T, MANGINO M, MARTEN J, MIHAILOV E, MILLER MB, VAN DER MOST PJ, OLDMEADOW C, PAYTON A, PERVJAKOVA N, PEYROT WJ, QIAN Y, RAITAKARI O, RUEEDI R, SALVI E, SCHMIDT B, SCHRAUT KE, et al. 2018. Gene discovery and polygenic prediction from a genome-wide association study of educational attainment in 1.1 million individuals. Nature Genetics, 50, 1112–1121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. LEVEY DF, POLIMANTI R, CHENG Z, ZHOU H, NUÑEZ YZ, JAIN S, HE F, SUN X, URSANO RJ, KESSLER RC, SMOLLER JW, STEIN MB, KRANZLER HR & GELERNTER J 2019. Genetic associations with suicide attempt severity and genetic overlap with major depression. Transl Psychiatry, 9, 22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. LYNGE E, SANDEGAARD JL & REBOLJ M 2011. The Danish National Patient Register. Scand J Public Health, 39, 30–3. [DOI] [PubMed] [Google Scholar]
  20. MARTIN AR, DALY MJ, ROBINSON EB, HYMAN SE & NEALE BM 2019. Predicting Polygenic Risk of Psychiatric Disorders. Biol Psychiatry, 86, 97–109. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. MORS O, PERTO GP & MORTENSEN PB 2011. The Danish Psychiatric Central Research Register. Scand J Public Health, 39, 54–7. [DOI] [PubMed] [Google Scholar]
  22. MORTHORST B, SOEGAARD B, NORDENTOFT M & ERLANGSEN A 2016. Incidence Rates of Deliberate Self-Harm in Denmark 1994–2011. Crisis, 37, 256–264. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. MULLINS N, BIGDELI TB, BØRGLUM AD, COLEMAN JRI, DEMONTIS D, MEHTA D, POWER RA, RIPKE S, STAHL EA, STARNAWSKA A, ANJORIN A, CORVIN A, SANDERS AR, FORSTNER AJ, REIF A, KOLLER AC, ŚWIĄTKOWSKA B, BAUNE BT, MÜLLER-MYHSOK B, PENNINX B, PATO C, ZAI C, RUJESCU D, HOUGAARD DM, QUESTED D, LEVINSON DF, BINDER EB, BYRNE EM, AGERBO E, STREIT F, MAYORAL F, BELLIVIER F, DEGENHARDT F, BREEN G, MORKEN G, TURECKI G, ROULEAU GA, GRABE HJ, VÖLZKE H, JONES I, GIEGLING I, AGARTZ I, MELLE I, LAWRENCE J, WALTERS JTR, STROHMAIER J, SHI J, HAUSER J, BIERNACKA JM, VINCENT JB, KELSOE J, STRAUSS JS, LISSOWSKA J, PIMM J, SMOLLER JW, GUZMAN-PARRA J, BERGER K, SCOTT LJ, JONES LA, AZEVEDO MH, TRZASKOWSKI M, KOGEVINAS M, RIETSCHEL M, BOKS M, ISING M, GRIGOROIU-SERBANESCU M, HAMSHERE ML, LEBOYER M, FRYE M, NÖTHEN MM, ALDA M, PREISIG M, NORDENTOFT M, BOEHNKE M, O’DONOVAN MC, OWEN MJ, PATO MT, RENTERIA ME, BUDDE M, WEISSMAN MM, WRAY NR, BASS N, CRADDOCK N, SMELAND OB, ANDREASSEN OA, MORS O, GEJMAN PV, SKLAR P, MCGRATH P, HOFFMANN P, MCGUFFIN P, LEE PH, MORTENSEN PB, KAHN RS, OPHOFF RA, ADOLFSSON R, VAN DER AUWERA S, DJUROVIC S, KLOIBER S, HEILMANN-HEIMBACH S, et al. 2019. GWAS of Suicide Attempt in Psychiatric Disorders and Association With Major Depression Polygenic Risk Scores. Am J Psychiatry, 176, 651–660. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. MURRAY GK, LIN T, AUSTIN J, MCGRATH JJ, HICKIE IB & WRAY NR 2021. Could Polygenic Risk Scores Be Useful in Psychiatry?: A Review. JAMA Psychiatry, 78, 210–219. [DOI] [PubMed] [Google Scholar]
  25. NORDENTOFT M, MORTENSEN PB & PEDERSEN CB 2011. Absolute risk of suicide after first hospital contact in mental disorder. Arch Gen Psychiatry, 68, 1058–64. [DOI] [PubMed] [Google Scholar]
  26. NØRGAARD-PEDERSEN B & HOUGAARD DM 2007. Storage policies and use of the Danish Newborn Screening Biobank. J Inherit Metab Dis, 30, 530–6. [DOI] [PubMed] [Google Scholar]
  27. O’CONNOR LJ & PRICE AL 2018. Distinguishing genetic correlation from causation across 52 diseases and complex traits. Nat Genet, 50, 1728–1734. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. ORRI M, SÉGUIN JR, CASTELLANOS-RYAN N, TREMBLAY RE, CÔTÉ SM, TURECKI G & GEOFFROY MC 2021. A genetically informed study on the association of cannabis, alcohol, and tobacco smoking with suicide attempt. Mol Psychiatry, 26, 5061–5070. [DOI] [PubMed] [Google Scholar]
  29. PEDERSEN CB 2011. The Danish Civil Registration System. Scand J Public Health, 39, 22–5. [DOI] [PubMed] [Google Scholar]
  30. PEDERSEN CB, BYBJERG-GRAUHOLM J, PEDERSEN MG, GROVE J, AGERBO E, BÆKVAD-HANSEN M, POULSEN JB, HANSEN CS, MCGRATH JJ, ALS TD, GOLDSTEIN JI, NEALE BM, DALY MJ, HOUGAARD DM, MORS O, NORDENTOFT M, BØRGLUM AD, WERGE T & MORTENSEN PB 2018. The iPSYCH2012 case-cohort sample: new directions for unravelling genetic and environmental architectures of severe mental disorders. Mol Psychiatry, 23, 6–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. PETERSSON F, BAADSGAARD M & THYGESEN LC 2011. Danish registers on personal labour market affiliation. Scand J Public Health, 39, 95–8. [DOI] [PubMed] [Google Scholar]
  32. PEYROT WJ, LEE SH, MILANESCHI Y, ABDELLAOUI A, BYRNE EM, ESKO T, DE GEUS EJ, HEMANI G, HOTTENGA JJ, KLOIBER S, LEVINSON DF, LUCAE S, MARTIN NG, MEDLAND SE, METSPALU A, MILANI L, NOETHEN MM, POTASH JB, RIETSCHEL M, RIETVELD CA, RIPKE S, SHI J, WILLEMSEN G, ZHU Z, BOOMSMA DI, WRAY NR & PENNINX BW 2015. The association between lower educational attainment and depression owing to shared genetic effects? Results in ~25,000 subjects. Mol Psychiatry, 20, 735–43. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. RUDERFER DM, WALSH CG, AGUIRRE MW, TANIGAWA Y, RIBEIRO JD, FRANKLIN JC & RIVAS MA 2020. Significant shared heritability underlies suicide attempt and clinically predicted probability of attempting suicide. Mol Psychiatry, 25, 2422–2430. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. SALVATORE JE, BARR PB, STEPHENSON M, ALIEV F, KUO SI, SU J, AGRAWAL A, ALMASY L, BIERUT L, BUCHOLZ K, CHAN G, EDENBERG HJ, JOHNSON EC, MCCUTCHEON VV, MEYERS JL, SCHUCKIT M, TISCHFIELD J, WETHERILL L & DICK DM 2020. Sibling comparisons elucidate the associations between educational attainment polygenic scores and alcohol, nicotine and cannabis. Addiction, 115, 337–346. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. SEARLE SR, SPEED FM & MILLIKEN GA 1980. Population Marginal Means in the Linear Model: An Alternative to Least Squares Means. The American Statistician, 34, 216–221. [Google Scholar]
  36. SELZAM S, KRAPOHL E, VON STUMM S, O’REILLY PF, RIMFELD K, KOVAS Y, DALE PS, LEE JJ & PLOMIN R 2017. Predicting educational achievement from DNA. Molecular Psychiatry, 22, 267–272. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. STRAWBRIDGE RJ, WARD J, FERGUSON A, GRAHAM N, SHAW RJ, CULLEN B, PEARSALL R, LYALL LM, JOHNSTON KJA, NIEDZWIEDZ CL, PELL JP, MACKAY D, MARTIN JL, LYALL DM, BAILEY MES & SMITH DJ 2019. Identification of novel genome-wide associations for suicidality in UK Biobank, genetic correlation with psychiatric disorders and polygenic association with completed suicide. EBioMedicine, 41, 517–525. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. SÖRBERG WALLIN A, ZEEBARI Z, LAGER A, GUNNELL D, ALLEBECK P & FALKSTEDT D 2018. Suicide attempt predicted by academic performance and childhood IQ: a cohort study of 26 000 children. Acta Psychiatr Scand, 137, 277–286. [DOI] [PubMed] [Google Scholar]
  39. SØRENSEN HJ, DEBOST JC, AGERBO E, BENROS ME, MCGRATH JJ, MORTENSEN PB, RANNING A, HJORTHØJ C, MORS O, NORDENTOFT M & PETERSEN L 2018. Polygenic Risk Scores, School Achievement, and Risk for Schizophrenia: A Danish Population-Based Study. Biol Psychiatry, 84, 684–691. [DOI] [PubMed] [Google Scholar]
  40. TAQUET M, SMITH SM, PROHL AK, PETERS JM, WARFIELD SK, SCHERRER B & HARRISON PJ 2021. A structural brain network of genetic vulnerability to psychiatric illness. Mol Psychiatry, 26, 2089–2100. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. TORKAMANI A, WINEINGER NE & TOPOL EJ 2018. The personal and clinical utility of polygenic risk scores. Nature Reviews Genetics, 19, 581–590. [DOI] [PubMed] [Google Scholar]
  42. VISSCHER PM, WRAY NR, ZHANG Q, SKLAR P, MCCARTHY MI, BROWN MA & YANG J 2017. 10 Years of GWAS Discovery: Biology, Function, and Translation. Am J Hum Genet, 101, 5–22. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. WORLD HEALTH, O. 1993. ICD-10 Classification of Mental and Behavioural Disorders : Diagnostic Criteria for Research, Albany, SWITZERLAND, World Health Organization. [Google Scholar]
  44. WRAY NR, LEE SH, MEHTA D, VINKHUYZEN AAE, DUDBRIDGE F & MIDDELDORP CM 2014. Research Review: Polygenic methods and their application to psychiatric traits. Journal of Child Psychology and Psychiatry, 55, 1068–1087. [DOI] [PubMed] [Google Scholar]
  45. WRAY NR, LIN T, AUSTIN J, MCGRATH JJ, HICKIE IB, MURRAY GK & VISSCHER PM 2021. From Basic Science to Clinical Application of Polygenic Risk Scores: A Primer. JAMA Psychiatry, 78, 101–109. [DOI] [PubMed] [Google Scholar]
  46. ZHU Z, ZHENG Z, ZHANG F, WU Y, TRZASKOWSKI M, MAIER R, ROBINSON MR, MCGRATH JJ, VISSCHER PM, WRAY NR & YANG J 2018. Causal associations between risk factors and common diseases inferred from GWAS summary data. Nat Commun, 9, 224. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Supp 1
NIHMS1862234-supplement-Supp_1.docx (23.6KB, docx)

Data Availability Statement

Due to safety regulations, data are not available to share.

RESOURCES