No association between war-related trauma or PTSD symptom severity and epigenome-wide DNA methylation in Burundian refugees

Figures & data

Table 1. Sociodemographic characteristics in children and caregivers.

	Children (n = 110)	Caregivers (n = 207)
Age in years, M (SD, range)	12.2 (2.1, 7–15)	37.4 (10.3, 19–74)
Education, n
No schooling	10	57
Class 1–3 (primary)	58	37
Class 4–6	41	74
Some secondary school years	1	28
Completed secondary school	–	11
Country of origin, n
Burundi	74	194
Tanzania	36	10
Democratic Republic of Congo	0	2
Other	0	1
Household size, M (SD, range)	7.2 (1.8, 3–12)	7.3 (1.9, 3–13)
Household income per month (USD), n
No income	–	69
Up to 20	–	115
More than 20	–	23

Note: n = conditional sample size, M = mean, SD = standard deviation, USD = US-Dollar, – = not applicable.

Table 2. Prevalence rates of mental health problems in the child and caregiver samples.

	Children (n = 110)		Caregivers (n = 207)
	Female (n = 54)	Male (n = 56)	Female (n = 108)	Male (n = 99)
Number of war-related traumatic event types, M (SD)	3.2 (3.0)	4.0 (4.0)	12.2 (4.5)	15.5 (4.1)
PTSD diagnosis, n	2	4	39	27
PTSD symptom severity, M (SD)	15.2 (11.8)	20.3 (16.4)	40.5 (19.2)	34.1 (16.3)
Substance use, M (SD)	–	–	7.7 (11.9)	17.4 (17.8)

Note: The caregivers’ measures significantly differed between the sexes (ps < .05), the children’s did not (ps > .05). n = conditional sample size, M = mean, SD = standard deviation, PTSD = post-traumatic stress disorder, – = not applicable.

Figure 1. Stacked manhattan plot of retrieved CpGs from EWAS for children. For the sake of illustration, −log10 transformed p-value of each CpG analysed in the two children’s EWAS including either the predictor of war-related trauma load or PTSD symptom severity are plotted in order of its location on the autosome. CpGs on the sex chromosomes were not analysed. The dashed line indicates the array-wide nominal threshold of −log10(9 × 10⁻⁸). There were no significant DMPs associated with any children’s predictor.

Figure 2. Stacked manhattan plot of retrieved CpGs from EWAS for caregivers. For the sake of illustration, −log10 transformed p-value of each CpG analysed in the two caregivers’ EWAS including either the predictor of war-related trauma load or PTSD symptom severity are plotted in order of its location on the autosome. CpGs on the sex chromosomes were not analysed. The dashed line indicates the array-wide nominal threshold of −log10(9 × 10⁻⁸). There were no significant DMPs associated with any caregivers’ predictor.

Data availability statement

DNAm and phenotypic data are not publicly available due to de-anonymization risk but available from the corresponding author on reasonable request.