Table 3.

The longitudinal association between nightmares and psychiatric symptoms.

Citation	Design	N	Sample	Measure of nightmares	Measure of psychiatric symptoms	Main findings	Depr.	PTSD	Suicidal ideation / behaviour	Psychotic experiences
Bernert et al. (2017)	A 21-day assessment period with three time periods (baseline, 7 and 21 days).	50	Students aged 18 and older with a suicide attempt history and recent suicidal ideation.	Nightmare severity assessed by the DDNSI.	Suicide (Beck scale for suicide ideation).	Nightmares associated with higher severity of suicidal ideation at 7 and 21 day follow up, after controlling for depression and baseline suicidal ideation. Insomnia and nightmares combined accounted for two thirds of the variance in suicidal ideation (R2 = 0.69 at 7 days and 0.59 at 21 days). Nightmares were not a significant predictor of later mood variability.	N/A	N/A	Yes	N/A
Fisher et al. (2014)	Birth cohort study. Children's nightmares were assessed between ages 2.5 and 9 years. Psychotic experiences.	6796	Children and their mothers.	Maternal report (“In the past year, has your child regularly had nightmares?”). Aged 12, children were asked “Since your 12th birthday have you had any dreams that woke you up? Were they frightening?”.	Key psychotic experiences including hallucinations, delusions, thought interference (the psychosis-like symptom interview).	Frequent childhood nightmares predicted later psychotic experience at 12 years, after controlling for sex, family adversity, emotional or behavioural problems, IQ and potential neurological problems (OR = 1.16, p < .05).	N/A	N/A	N/A	Yes
Gerhart et al. (2014)	Assessed at baseline and 6 months later.	779	Palestinian adults (amid ongoing violent political turmoil)	1-Item of the PSS assessed trauma related nightmares.	PTSD (PSS), depression (PHQ-9).	Trauma related nightmares predicted later PTSD (β = 0.10, p < .05) and depression (β = 0.10, p < .05) but the reverse relationships were not significant.	Yes	Yes	N/A	N/A
Greene, Gregory, Fone, & White (2015)	Assessed at 5 years of age (parent report) and assessed for risk for depression at 34 years.	7437	Children and their parents.	Parental reports on the frequency of nightmares.	Self-reported depression (yes/no) and if yes, whether this was within the past year.	Parent reports of childhood nightmares at age 5 were not associated with depression over the past year at age 34 (OR: 1.03, 95% CI: 0.84–1.56).	No	N/A	N/A	N/A
Hedström et al. (2020)	Baseline and on average 16 years later	40,902 (69 suicides within the sample)	Swedish national march cohort	Nightmare frequency (dichotomous: never or seldom versus sometimes to always)	Depression (diagnosis from patient register), depressive symptoms (self reported feeling sad, low-spirited or depressed). Suicide (death register, 69 suicides registered within the sample).	Often or always having nightmares was associated with a significantly increased incidence of suicide. However, after adjustment (for seven covariates) statistical significance was lost. Among participants without a diagnosis of depression at baseline, the odds of depression during follow-up was higher among those who suffered from nightmares than among those who did not (OR 1.35, 95% CI 1.19–1.53).	Yes	N/A	Yes	N/A
Kobayashi & Delahanty (2013)	Sleep assessed two weeks after a trauma and PTSD at seven weeks	45 split by gender (female = 17, male = 28).	Participants admitted to a trauma centre	Nightmares assessed by the PSQI-addendum for PTSD	PTSD (CAPS)	Nightmares were associated with later PTSD in men (r = 0.53, p < .01) but the relationship was not significant in women (r = 0.30, p > .05).	N/A	Mixed	N/A	N/A
Lereya et al. (2017)	Sleep assessed when children were 2.5, 3.5, 4.8 and 6.8 years of age. Borderline personality disorder symptoms assessed age 11–12.	5544	Adolescents	Mother questionnaire report of nightmares (“In the past year has your child regularly had nightmares?” yes any sleep problem versus no sleep problem)	Borderline personality disorder (UK childhood interview for DSM-IV borderline personality disorder)	Having persistent childhood nightmares was associated with later borderline personality disorder symptoms in adolescence after controlling for sleep onset and maintenance problems and confounders including psychiatric diagnosis, emotional and behavioural problems, abuse and family adversity (adjusted OR = 1.62; 95% CI = 1.12–2.32).	N/A	N/A	N/A	N/A
Li et al. (2010)	Outpatients completed a comprehensive sleep assessment (alongside their clinical assessment). Medical notes were subsequently reviewed one year later.	1231	Psychiatric outpatients aged 18–65 from Hong Kong	Nightmare frequency assessed over the past year, determined on a 5-point scale (0 = none, to 4 = >3 times per week).	A suicide attempt was defined as an act of intentional self-harm to end one's life, as documented by case notes.	Recurrent nightmares predicted later suicide attempts (OR = 8.17, 95% CI, 1.06–63.13). The addition of comorbid insomnia to nightmares increased the one-year incidence of suicide risk (OR = 17.08; 95% CI, 2.64–110.40).	N/A	N/A	Yes	N/A
Li et al. (2012)	4-year naturalistic follow-up. Analysis controlled for age, marital status, psychiatric comorbidity, and severity of depressive and anxiety symptoms.	371	Remitted depressed outpatients (mean age 44.6 ± 10.4 years, female 81.8%)	Sleep questionnaire Nightmare Distress Questionnaire	Depression and anxiety (MINI at baseline, HADS at follow up), suicidality (MINI suicidality module).	Participants experiencing at least one nightmare per week at baseline were less likely to be remitted from depression at follow-up compared with those without nightmares (with nightmares: 29.8% versus without: 47.3%, p < .01).	Yes	N/A	Yes	N/A
Li et al. (2016)	Baseline and approximately one year later.	388	Outpatients diagnosed with schizophrenia spectrum disorders.	Frequent nightmares were defined as having nightmares of at least once per week over the past year	Suicide attempt recorded in medical notes. A suicide attempt was defined as a deliberate act of self-harm with the intention to end one's life.	Nightmare complaint alone did not predict the occurrence of suicide attempts, but the comorbidity of nightmares and insomnia was associated with the risk of suicide attempt over follow-up (adjusted HR = 11.10, p < .05).	N/A	N/A	No	N/A
Liu et al. (2019)	Baseline and 1 year follow up	7072	Adolescents	One survey question: in the past year how often did you have nightmares? Response option was 7 point likert scale from never to almost every night.	Silverman et al. (2007) suicidal behaviour measure.	Frequent nightmares at baseline were significantly associated with suicide attempts and non-suicidal self-injury one year later. After adjusting for demographics, depression, impulsivity and prior suicide attempt, the association remained significant for suicide attempts (OR = 1.96, 95% CI = 1.15–3.33) and non-suicidal self-injury (OR = 1.1.52, 95% CI = 1.10–2.08). Adjustment for insomnia and sleep duration yielded almost no change.	N/A	N/A	Yes	N/A
Liu et al. (2020)	Baseline and 1 year follow up.	6923	Adolescents	One survey question: “During the past month, how many times did you have nightmares?” The participant entered number of nights with nightmares.	Suicidal thoughts: “Have you ever seriously thought about suicide or killing yourself over the past 12 months?” (yes/no). Suicidal plans:”“Have you ever had a specific plan for how you would kill yourself over the past 12 months?” (yes/no), suicide attempt: “Have you ever tried to kill yourself over the past 12 months?” (yes/no). Depression (CES—D).	This is the same sample as Liu et al. (2019). The log odds of endorsing suicidal thoughts, plans or attempts at follow-up were all significantly higher in those endorsing frequent nightmares at baseline. The relationship was partially mediated by depression. Adjusting for covariates and baseline suicidal behaviour reduced the size of effects. The same set of analysis assessing the relationship between nightmares distress and suicidal behaviour yielded mixed results.	Yes	N/A	Yes	N/A
Neylan et al. (2020)	Baseline (in emergency department) and 2 and 8 weeks follow up.	666	Adults experiencing a motor vehicle collision	Two modified questions from the CAPS for DSM-IV assessing frequency and distress caused by unpleasant dreams. Questions asked about bad dreams in general in the 30 days leading up to the accident, rather than linked to a specific trauma.	PTSD and acute stress disorder (PTSD Checklist for DSM-5; PCL-5), Peritraumatic distress and dissociation (8-item short-form of the 13-item Peritraumatic Distress Inventory and the 5-item revised Michigan Critical Events Perception Scale), depression (PROMIS short form).	Nightmares were a significant predictor of 2 week acute stress disorder (ASD) and 8 week PTSD, whilst controlling for retrospective reports of pre-trauma PTSD and peritraumatic symptoms (2 week ASD OR = 1.3, 95% CI = 1.0–1.5, 8 week PTSD OR = 1.3, 95% CI = 1.1–1.6). Nightmares predicted major depressive episode at 8 weeks (OR = 1.3, 95% CI = 1.0–1.6), whilst controlling for pre-trauma PTSD and depression. The association held whilst controlling for peritraumatic symptoms (OR = 1.2, 95% CI = 1.0–1.5). Nightmares were not significantly associated with depression at two weeks post-trauma (OR = 1.2, 95% CI = 1.0–1.5).	Yes	Yes	N/A	N/A
Pigeon, Campbell, Possemato, & Ouimette (2013)	Assessed at baseline and 6 months following baseline.	80	Combat veterans with hazardous alcohol use and at least subthreshold PTSD.	1-Item of PTSD Checklist – Military Version Coded as positive if “moderately bothered” by trauma related nightmares.	PTSD (The PTSD checklist), depression (CES—D).	Bothersome trauma related nightmares predict later PTSD severity. Of those endorsing nightmares at baseline, 41% developed PTSD, compared with 10% of those without. Trauma related nightmares were not associated with later depression.	No	Yes	N/A	N/A
Sandman et al. (2017)	Finnish National FINRISK Study assessed between 1972 and 2012 and followed up in 2014 or at point of death.	71,068	i) General population of Finland and ii) war veterans.	Survey question asks: “During the past 30 days have you had nightmares?”, (“often”, “sometimes” or “never.”).	Cases of suicide were measured using the Causes of Death Registry of Finland.	The unadjusted hazard ratio of suicide for persons who reported frequent nightmares compared with those who did not was 2.63 (p < .001), after adjustments this reduced to 1.84 (p = .010). The presence or absence of war veterans did not affect the association.	N/A	N/A	Yes	N/A
Shi et al. (2021)	Baseline and 6 months later	11,740	College students	Survey question: How often did you have nightmares during the past month (1 = never, 4 = more than three per week).	Survey questions: “Have you ever seriously thought about suicide during the past 12 months”, “Have you ever tried to kill yourself during the past 12 months” (yes/no). Similar questions were used to assess suicidal behaviours over the past 6 months at follow-up.	Endorsing frequent nightmares was associated with later suicidal ideation (OR: 1.69) and attempts (OR: 2.40) after controlling for a range of covariates including depression and baseline measure of the outcome.	N/A	N/A	Yes	N/A
Sjöström et al. (2009)	Assessed nightmares at baseline and two months follow up. Repeat suicide attempts assessed at 2 years follow-up.	165	Psychiatric inpatients admitted for suicide attempt aged between 18 and 69 years.	1-item of the Uppsala Sleep Inventory answered on a five-point frequency scale (1 = never, to 5 = very often). Score ≥ 4 at both baseline and follow-up was considered a persistent sleep disturbance.	Suicide attempt data taken from medical records. Psychiatric symptoms (SCID for DSM-IV and the Comprehensive Psychopathological Self-Rating Scale for Affective Syndromes).	Having frequent nightmares at baseline predicted repeat suicide attempts (OR = 3.15, 95% CI, 1.51–6.57). Persistent nightmares further heightened the risk (OR = 5.20, 95% CI, 1.91–14.13). Results remain even after adjusting for sex, axis-I diagnosis, depression, anxiety, PTSD and anti-depressant drugs. Decrease in odds ratio (OR = 2.28) after controlling for depression.	N/A	N/A	Yes	N/A
Tanskanen et al. (2001)	Finnish National FINRISK Study assessed between 1972 and 1992 and followed until December 31, 1995, or death	36,211	General population of Finland.	Survey question asks: ““How often have you had nightmares during the past month?” (“frequently,” “occasionally,” or “not at all.”).	Cases of suicide were measured using the Causes of Death Registry of Finland	Significant association between nightmares and later death by suicide. Among subjects having nightmares occasionally the adjusted relative risk of suicide was 57% higher, and among those reporting frequent nightmares 105% higher compared with subjects reporting no nightmares at all.	N/A	N/A	Yes	N/A
Thompson et al. (2015)	Follow up from Fisher et al. (2014). Children were interviewed at ages 12 and 18 years about nightmares and psychotic experiences.	4720	Children and their mothers.	Postal questionnaires completed by mothers when children aged 2.5–9 (“In the past year, has your child regularly had nightmares?”). At aged 12 children were asked “Since your 12th birthday have you had any dreams that woke you up? Were they frightening?”.	Key psychotic experiences including hallucinations, delusions, thought interference (the psychosis-like symptom interview).	Nightmares at 12 was associated with psychotic experiences age 18 when adjusting for IQ, family adversity, psychiatric disorders, depression, child abuse enuresis and psychotic experiences at 12 (OR = 1.62, 95% CI 1.19–2.20). The effect was larger for persistent psychotic experiences (at age 12 and 18 years, OR: 3.87 95% CI: 2.30–6.51).	N/A	N/A	N/A	Yes
van Liempt et al. (2013)	Assessed prior to military deployment to Afghanistan and at 6 months post-deployment.	453	Dutch service members (military)	1-Item of the Self-Rating Inventory for PTSD (“I had bad dreams”).	PTSD (SRIP).	Pre-deployment nightmares predicted later PTSD symptoms at six months post-deployment (OR = 2.99, 95% CI: 1.10–8.55), after controlling for pre-deployment PTSD, early trauma, mood and anxiety.	N/A	Yes	N/A	N/A
Wittmann et al. (2010)	Assessed at 10 days, 2 months, and 6 months after a traffic accident.	32	Children who experienced traffic accidents (Age 11.9; SD = 2.4; range = 7–15 years)	Trauma nightmare frequency scale of Clinician-Administered PTSD Scale, child and adolescent version, 1 item.	PTSD (CAPS child and adolescent version). Depression (children's depression inventory).	Trauma related nightmares ten days post-accident predict later PTSD total score two months post-accident, but not depression.	No	Yes	N/A	N/A
Wong et al. (2011)	Sleep problems assessed age 12–14. Suicidal thoughts and self-harm assessed age 15–17.	391	Children of parents who are alcoholics	One question on the Youth Self Report	Two items on the Youth Self Report.	Nightmares at age 12–14 did not predict later suicidal ideation or self-harm when ‘trouble sleeping’ was included in the model.	N/A	N/A	No	N/A

Anx. = anxiety. Depr = depression. PTSD = post-traumatic stress disorder.