Table 3.

Study findings and synthesis

Study	Primary results				Mediation/moderation analyses	Main limitations
	Prenatal depression	Postnatal depression (birth to 12 months)	Beyond perinatal depression (after 12 months)	Recurrence/chronicity/cumulative exposure
Barker et al. [64]	Prenatal MD independently predicted offspring behavioural problems after adjustment for prenatal anxiety, later MD and covariates. No interaction between offspring gender and MD	N/A	Later MD (1.5 years) predicted offspring behavioural problems independently (of prenatal) predicted externalising. No interaction between offspring gender and MD	N/A	N/A	Sample: Low rates of ethnic minorities. Substantial attrition over time Measures: Maternal report of depression and child outcome. Depression not measured concurrently Other: Only included 1 assessment of later maternal depression
Brennan et al. 2000 [49]	Prenatal MD did not predict offspring behavioural problems after adjustment for covariates and MD at other time points. No evidence of gender differences	Postnatal MD predicts elevated offspring total problems after adjustment for covariates and MD at other time points. No evidence of gender differences	Concurrent MD predicts elevated offspring total problems after adjustment for covariates and MD at other time points. No evidence of gender differences	Chronicity and severity of MD predicts elevated offspring behavioural problems after adjustment for covariates and MD at other time points. No evidence of gender differences	N/A	Sample: Lower SES than the population. Non-random attrition (those lost differed in a number of ways from those retained) Measures: Maternal report of depression and child outcome. Other: Child outcome did not distinguish between behavioural and emotional problems. Only included 1 assessment of later maternal depression (concurrent)
Edwards and Hans [43]	The initial association between prenatal MD and offspring behavioural problems was mediated by maternal sensitivity and concurrent maternal MD. Effects only significant for boys	N/A	The initial association between prenatal MD and offspring behavioural problems was mediated by maternal sensitivity and concurrent maternal MD. Effects only significant for boys	N/A	Maternal sensitivity mediates the relationship between prenatal MD and offspring behavioural problems	Sample: Small homogenous sample not representative of general population Measures: Maternal report of depression and child outcome. Other: Outcome did not distinguish between behavioural/emotional problems. Only included 1 assessment of later maternal depression (concurrent)
Eichler et al. [65]	Prenatal MD predicts offspring antisocial behaviour after controlling for current MD and covariates. Effects significant for boys but not girls	N/A	Concurrent MD (child 6-9 years) did not predict offspring antisocial behaviour after adjustment for covariates and prenatal MD	N/A	N/A	Sample: Significantly higher education, makes generalisation difficult. Relatively small sample size Measures: Mother self-report on depression and child outcome Other: Only included 1 assessment of later maternal depression (concurrent)
Gjerde et al. [61]	Prenatal MD does not predict offspring behavioural problems after adjustment for unmeasured confounding. Gender differences not tested	Postnatal MD does not predict offspring behavioural problems after adjustment for unmeasured confounding	Concurrent MD predicts elevated offspring behavioural problems after adjustment for unmeasured confounding	N/A	Unmeasured familial confounding explains initial association between prenatal MD and offspring behavioural problems	Sample: Substantial attrition Measures: Mother self-report on depression and child outcome Other: Only included 1 assessment of later maternal depression (concurrent)
Hanington, et al. [44]	Prenatal MD (2nd trimester) and marital conflict during pregnancy predicted elevated offspring conduct problems when postnatal MD and paternal pre-and-postnatal depression were accounted for. Gender differences not tested	Postnatal MD (8 months) predicted elevated offspring conduct problems when prenatal MD, paternal pre-and-postnatal depression, and marital conflict were accounted for	N/A	N/A	Marital conflict partially mediated the relationship between postnatal MD and offspring conduct problems	Sample: Low rates of ethnic minorities. Substantial attrition over time Measures: Mother self-report on depression and child outcome Other: Does not include measures of later maternal depression
Hay et al. (2003) [51]	Prenatal MD did not predict offspring violence after subsequent episodes of MD and covariates were adjusted for. Boys more likely to show violence than girls	Postnatal MD predicted offspring violence after adjustment for prenatal MD, later MD and covariates. Boys more likely to show violence than girls	Current psychological distress (GAS Score) associated with child violence but did not explain the impact of postnatal MD	Risk for violence was greatest in the group of children whose mothers were depressed in the postpartum period and at least once thereafter	N/A	Sample: Small sample size drawn from a large metropolitan area and more ethnically diverse than the general population Measures: Retrospective reports of maternal depression between developmental time periods
Hay et al. (2010) [48]	Prenatal MD significantly predicted offspring violence after controlling for covariates and exposure to subsequent episodes of MD. Boys and girls were not significantly different in ASB or violence specifically	Postnatal MD did not predict offspring violence after covariates and maternal MD at other time points was taken into account. Boys and girls were not significantly different in ASB or violence specifically	Exposure to MD during either early, middle or late childhood did not predict offspring ASB once the impact of prenatal MD was accounted for	Recurrent exposure to maternal MD predicted offspring violence but did not explain the effect of prenatal depression	N/A	Sample: Small sample size drawn from a large metropolitan area and more ethnically diverse than the general population Measures: Retrospective reports of maternal depression between developmental time periods
Hay et al. (2008) [47]	Prenatal MD did not predict elevated risk of offspring disruptive behavioural disorders. No significant differences between boys and girls in risk of disruptive behavioural disorders	Postnatal MD did not predict elevated risk of offspring disruptive behavioural disorders. No significant differences between boys and girls in risk of disruptive behavioural disorders	Exposure to maternal depression after 3 months postpartum predicted elevated rates of offspring DBD. Exposure to later MD not analysed at the level of specific time points	Cumulative exposure to maternal depression after the perinatal period (during early to late childhood and/or adolescence) predicted elevated rates of offspring DBD	N/A	Sample: Small sample size drawn from a large metropolitan area and more ethnically diverse than the general population Measures: Retrospective reports of maternal depression between developmental time periods
Korhonen et al. (2012) [70]	Prenatal MD (3rd trimester) predicted offspring reported behavioural problems following adjustment for concurrent MD. Findings only significant for boys. Prenatal MD did not predict maternal reports of offspring behavioural problems	Postnatal MD predicted offspring reported behavioural problems following adjustment for concurrent MD. Findings only significant for male offspring. Postnatal MD did not predict maternal reports of offspring behavioural problems	Concurrent MD predicted elevated offspring behavioural problems according to mothers and adolescents own self-reports. Findings significant for males only	Recurrent (number of times scored over the EPDS cut point) MD predicted elevated offspring behavioural problems. Gender differences not tested due to small cell sizes	N/A	Sample: Number of symptomatic mothers and children was low; Attrition was high. Relatively moderate sample size Measure: Limited control for sociodemographic and other risk factors Other: Only included 1 assessment of later maternal depression (concurrent)
Korhonen et al. (2014) [69]	Prenatal MD predicted elevated offspring behavioural problems prior to controlling for other episodes. Results only significant for mothers’ but not offspring self-reports. Gender differences not tested	Initial exposure to MD at 2 weeks or 6 months postpartum was not associated with elevated offspring behavioural problems for either informant. Gender differences not tested	For both informants, initial exposure to MD at 4–5 and 8–9 years was not associated with offspring behavioural problems. Prior to controlling for other episodes, initial exposure to MD at aged 16–17 (concurrent) years was associated with elevated offspring behavioural problems for self but not maternal reports	Chronic maternal depressive symptoms rather than exposure at specific time periods predict adolescent behavioural problems	N/A	Sample: Number of symptomatic mothers and children was low, attrition is high. Relatively moderate sample size Measure: Self-report measures of depression. Initial exposure to MD was used as a measure (the first time at each time point that a mother exceeded cut off for EPDS, rather than the independent effect)
Lahti et al. [66]	Prenatal MD had a direct effect on offspring externalising behavioural problems. Gender differences not tested	N/A	MD symptoms at the time of the child assessment (1.9 to 5.9 years) partially mediated the effect of prenatal MD on offspring behavioural problems	N/A	N/A	Sample: Homogenous sample, significant attrition Measures: Mother self-report on depression and child outcome Other: Only included 1 assessment of later maternal depression (concurrent)
Leis et al. [71]	Prenatal MD predicted an increase in total offspring conduct problems. Effect remained after controlling for later MD and anxiety for mothers but not teachers’ reports of child behavioural	Postnatal MD predicted increased offspring conduct problems for both mothers and teachers reports at the univariate level. The impact of prenatal depression and other covariates were not examined	N/A	N/A	N/A	Sample: High attrition Measures: Mother self-report on depression and child behaviour Other: The impact of post-pregnancy MD and the specific time periods that MD was measured were not reported/analysed in sufficient detail to extract the required information
Luoma et al. [67]	Prenatal MD predicted elevated offspring behavioural problems for combined maternal and teacher reports. Results remained significant after controlling for subsequent exposure to MD and covariates. Effects stronger for boys than girls	Postnatal MD did not predict offspring behavioural problems	Concurrent MD did not predict offspring behavioural problems. Effects stronger for boys than girls	N/A	N/A	Sample: Moderate-sized sample, small group sizes for MD Measures: Mother self-report on depression and child outcomes Other: Results are not clearly reported. Only included 1 assessment of later maternal depression (concurrent)
O’Connor et al. [62]	Prenatal MD does not predict offspring behavioural problems after the influence of prenatal anxiety is accounted for. No gender differences	Postnatal depression at 8-weeks and 8-months postpartum predicts offspring behavioural problems after controlling for prenatal depression and anxiety and covariates. No gender differences	N/A	N/A	N/A	Sample: Substantial Attrition Measures: All data based on maternal reports Other: Does not distinguish between child emotional/behavioural problems. Did not include an assessment of later maternal depression
O’Donnell et al. [72]	Prenatal MD predicted elevated offspring behavioural/emotional problems. The effects were observed after controlling for multiple confounders including prenatal anxiety and postnatal depression. No interaction between MD exposure and offspring gender	Unobtainable	Unobtainable	N/A	N/A	Measures: All data based on mother reports Other: Results were not reported in sufficient detail to extract the required information e.g. cannot distinguish post-pregnancy depression data. Did not distinguish child behavioural and emotional problems
Raskin et al. [60]	Prenatal MD predicted elevated offspring behavioural problems following adjustment for later MD and covariates. Gender differences not tested	N/A	MD between 1 and 2 years predicted elevated offspring behavioural problems following adjustment for prenatal MD and covariates. Gender differences not tested	Chronic MD from pregnancy to 2 years predicted elevated offspring behavioural problems. Gender differences not tested	N/A	Sample: High risk sample accessing a home visiting programme which limits generalizability Measures: Mother self-report on depression and child outcome. Potential shared method variance Other: Did not distinguish between child behavioural/emotional problems
Soe et al. [63]	Prenatal MD predicted elevated offspring behavioural problems following adjustment for covariates. Findings only significant for females	Postnatal MD predicted elevated offspring behavioural problems following adjustment for covariates. Findings only significant for females	N/A	N/A	N/A	Sample: High attrition Measures: Mother self-report on depression and child outcome. Potential shared method variance Other: Did not include a measure of later MD
van der Waerden et al. [68]	Prenatal MD did not predict elevated offspring behavioural problems following adjustment for covariates and later MD. Gender differences not reported	Postnatal MD predicted elevated offspring behavioural problems following adjustment for covariates and concurrent MD at 5 years. Gender differences not reported	Exposure to MD during the preschool years predicted elevated offspring behavioural problems following adjustment for covariates and concurrent MD at 5 years. Gender differences not eported	Exposure to chronic (persistent) MD from infancy and throughout early childhood predicted elevated rates of offspring behavioural problems. Gender differences not reported	N/A	Sample: Attrition not random (women depressed during pregnancy more likely to drop out) Measures: All data based on mother reports Other: Concurrent MD included as a covariate did not explore/report independent effect
Woolhouse et al. [9]	Prenatal MD predicts elevated offspring behavioural/emotional problems at the univariable level. Only combined exposure to prenatal and postnatal MD significant after adjustment for covariates. Boys had greater behavioural/emotional problems than girls	Postnatal MD predicts elevated offspring behavioural/emotional problems at the univariable level. Only combined exposure to postnatal and prenatal MD significant after adjustment for covariates. Boys had greater behavioural/emotional problems than girls	MD (concurrent) at 4 years postpartum had an independent effect on child behavioural/emotional problems after adjusting for prenatal and postnatal MD and covariates. Boys had greater behavioural/emotional problems than girls	N/A	N/A	Sample: Young women, single women and non-English speaking women are underrepresented. Selective attrition Measures: Maternal self-report of depression and child outcome Other: Did not distinguish between child behavioural/emotional problems. Only included 1 measure of later MD (concurrent)