Heritability of Perinatal Depression and Genetic Overlap With Nonperinatal Depression
Publication: American Journal of Psychiatry
Abstract
Objective:
The authors investigated the relative importance of genetic and environmental influences on perinatal depression, and the genetic overlap between perinatal depression and nonperinatal depression.
Method:
Analyses were conducted using structural equation modeling for 1) the lifetime version of the Edinburgh Postnatal Depression Scale in 3,427 Swedish female twins and 2) clinical diagnoses of depression separated into perinatal depression and nonperinatal depression in a Swedish population-based cohort of 580,006 sisters.
Results:
In the twin study, the heritability of perinatal depression was estimated at 54% (95% CI=35%−70%), with the remaining variance attributable to nonshared environment (46%; 95% CI=31%−65%). In the sibling design, the heritability of perinatal depression was estimated at 44% (95% CI=35%−52%) and the heritability of nonperinatal depression at 32% (95% CI=24%−41%). Bivariate analysis showed that 14% of the total variance (or 33% of the genetic variance) in perinatal depression was unique for perinatal depression.
Conclusions:
The heritability of perinatal depression was estimated at 54% and 44%, respectively, in separate samples, and the heritability of nonperinatal depression at 32%. One-third of the genetic contribution was unique to perinatal depression and not shared with nonperinatal depression, suggesting only partially overlapping genetic etiologies for perinatal depression and nonperinatal depression. The authors suggest that perinatal depression constitutes a subset of depression that could be prioritized for genomic discovery efforts. The study findings have direct translational impact that can assist clinicians in the counseling of their patients regarding risk and prognosis of perinatal depression.
Perinatal depression, defined as depressive illness occurring during pregnancy (antenatal depression) or following childbirth (postpartum depression), affects some 10%–15% of women and confers substantial morbidity, mortality, and personal and societal costs (1–4). The clinical presentation of perinatal depression features low mood, anxiety, rumination, and, in severe cases, suicidal or infanticidal ideation (5). Historically, perinatal depression has been conspicuously understudied compared with major depressive disorder (6).
Major depressive disorder is defined as marked and persistent depressed mood associated with physical and cognitive signs and symptoms. Depression is common and costly, and it is projected to be the second leading cause of disability worldwide by 2020 (7–10). The heritability of major depression has been estimated at 31%−42% (11, 12). In comparison with other major psychiatric disorders, discerning the genetic basis of depression has proven to be more challenging. Genome-wide linkage studies, candidate gene studies, and genome-wide association studies have not been successful in identifying risk loci that meet contemporary standards for replication (13). The relatively modest heritability of depression, as compared with other major psychiatric disorders, may be one reason for the substantially lower yield of identified genetic loci (14). Another reason is that depression is a markedly heterogeneous disorder.
In contrast, perinatal depression may represent a more homogeneous disorder. Perinatal depression occurs in women of reproductive age and is coupled with pregnancy and childbirth. The limited literature on the genetic basis of perinatal depression suggests a heritable component that may be greater than that in major depressive disorder (15–17). Two small studies have shown clustering in families (16, 17). Murphy-Eberenz et al. (16) reported odds ratios for prediction of sibling status for perinatal depression or postpartum depression between 2.28–3.96. Forty et al. (17) studied female siblings and found the sibling correlation for postpartum depression to be highest for postpartum depression with an onset within 6–8 weeks following childbirth (tetrachoric correlation coefficient=0.62, 95% CI=0.16–0.88; p=0.01) in women with recurrent depression. An Australian study of 1,676 twins (15) estimated the heritability of lifetime postpartum depression at 25%.
Some have argued that perinatal depression is partly or wholly distinct from major depressive disorder. According to this view, the biological underpinnings of perinatal depression differ from those of nonperinatal depression in that sensitivity to the dramatic fluctuations in gonadal hormone serum concentrations during the perinatal period probably play a pathogenic role (18, 19). This implies that perinatal depression and nonperinatal depression are at least partially different disorders, with perinatal depression featuring distinctive genetic and environmental etiological risk factors. The alternative viewpoint is that perinatal depression is merely an episode of major depressive disorder occurring sometime during the period of pregnancy or the immediate postpartum.
Given the uncertainty about the degree to which perinatal depression and nonperinatal depression are distinct and the limited literature on the genetic basis of depression during the perinatal period, there is a great need for an improved understanding of the genetic basis of perinatal depression and the extent of the genetic overlap between perinatal and nonperinatal depression. In this study, we used data from a validated screening tool for perinatal depression in 3,427 twins from the Swedish Twin Registry to estimate the relative contributions of genetic (heritability) and environmental risk factors to the liability to perinatal depression in a classical twin design. We then used Swedish population data from over 580,000 sisters and national treatment registers to estimate the heritability of perinatal depression, the heritability of nonperinatal depression, and the genetic and environmental overlap between the two.
Method
Classical Twin Study
Study population.
The Swedish Twin Registry contains almost 200,000 Swedish twins born between 1886 and 2008 (20). The substudy “Screening Across the Lifespan Twin Study: The Younger” (SALTY), conducted in 2009 and 2010, included 11,372 twins from the Swedish Twin Registry with a median birth year of 1950 (54.3% female). The SALTY study included an extensive self-report questionnaire that covered many different areas, including perinatal depression (21). The sample consisted of women from the SALTY study who reported having given birth to a living child and who completed the lifetime version of the Edinburgh Postnatal Depression Scale. This subgroup included 3,427 individual twins (1,516 female monozygotic and 1,911 female same-sex dizygotic twins) and both members of 1,106 twin pairs. Zygosity was determined using DNA for 27% of the twin females. For the remainder, zygosity was assigned based on questions about intrapair physical similarities in childhood (20).
Perinatal depression classification.
We assessed onset of mood symptoms both during pregnancy and the postpartum period using a retrospective lifetime version of the Edinburgh Postnatal Depression Scale (22). The Edinburgh Postnatal Depression Scale is the most widely used patient-rated assessment instrument for perinatal depressive illness in the world and has demonstrated good sensitivity and specificity in both antenatal and postpartum depression (2, 23). The lifetime version of the instrument includes the same 10 items used in the original scale but was modified to assess previously experienced (or lifetime) perinatal depression (22). A score ≥12 on the scale is the accepted standard cutoff to identify depressive illness and was used in this study to define a binary outcome (23, 24).
Statistical analysis.
The classical twin methodology relies on the different relatedness between monozygotic and dizygotic twins. Monozygotic twins are considered genetically identical, whereas dizygotic twins share an average of 50% of their segregating alleles. If genes influence a trait, there will be more pronounced twin similarity within monozygotic than within dizygotic pairs. By modeling twin covariance structures in monozygotic and dizygotic pairs, the variation in a phenotype is decomposed into additive genetic (A), shared environmental (C), and nonshared environmental (E) factors (25). We used a liability-threshold approach, assuming that the observed binary variable came from an underlying continuous liability of the trait (25). A threshold was assumed, and a 1 was assigned if an individual had a liability greater than the threshold, and 0 otherwise. The underlying liabilities were assumed to have normal distributions, and the correlations between these underlying normal distributions could be estimated (26). The resulting tetrachoric correlations form the basis of the heritability analysis. Note that the key assumptions of normally distributed liability and equal environments have strong empirical support (27).
Sibling Design
Study population.
To evaluate genetic and environmental influences on perinatal and nonperinatal depression in a larger and more generalizable setting, we included a population-based cohort from Swedish national register data. We used the Swedish personal identification numbers to link national Swedish longitudinal registries with high accuracy. The Swedish Medical Birth Register covers 99% of all births since 1973 (28). It was linked to the Multi-Generation Register, which contains information about first-degree relatives for persons born in 1932 and later (29). The Medical Birth Register and the Multi-Generation Register were linked to the Swedish Twin Registry to obtain information on twins and their zygosity. All parous women who had given birth to their first child after 1973 were included. The women also had to have been born in Sweden, could not have emigrated and moved back to Sweden more than once, and had to have at least one sister fulfilling the same criteria. Because of the low observed occurrence of perinatal depression in the registers (0.6%), we opted for a design that included up to four full or half-siblings per nuclear family, covering 99.8% of the eligible population. We identified a total of 580,006 parous female siblings from 260,384 unique families. This allowed for comparisons in 313,632 full-sister pairs, 28,568 maternal half-sister pairs, 33,931 paternal half-sister pairs, 2,104 dizygotic twin sister pairs, and 2,225 monozygotic twin sister pairs. A total of 1,572 twin sisters overlapped between the two different designs.
Disease classification.
We linked all subjects in the study population to the Swedish National Patient Register, which contains all Swedish psychiatric inpatient admissions since 1973 and psychiatric specialist outpatient treatment contacts since 2001 (30). The register contains admission dates along with the main discharge ICD diagnosis code and up to eight secondary diagnosis codes. Treatment contacts for depression were defined using ICD-8 codes 296.00, 296.40, 296.41, and 790.20; ICD-9 codes 296.2, 296.3, 296.9, 298.0, 300.4, 309.0, 309.1, and 311; and ICD-10 codes F32.0, F32.1, F32.2, F32.3, F32.8, F32.9, F33.0, F33.1, F33.2, F33.3, F33.4, F33.8, F33.9, F34.1, and F41.2. These diagnostic codes were selected to capture as many women with perinatal depressive symptoms as possible. The perinatal period was defined as any point from estimated date of conception through 6 months postpartum. Although the onset of postpartum depression is often within 4–6 weeks of childbirth, we intentionally expanded our definition to include women seeking care up to 6 months postpartum, as treatment onset may be substantially delayed after symptom onset. Many women delay seeking treatment because of concerns about the stigma of mental health treatment (31), uncertainty about the nature of their symptoms (postpartum “blues” or a normal transition to motherhood versus an illness state requiring treatment), or perceived lack of time for self-care (32).
The conception date was calculated using the birthdate of the child and gestational age at delivery. Perinatal depression was defined as ≥1 inpatient or outpatient treatment contact for unipolar depression within a perinatal period. Nonperinatal depression was defined as unipolar depression at any other time in a woman’s life.
The perinatal period was further separated into an antenatal and a postnatal period to allow heritability estimation of antenatal and postnatal depression, respectively. For this purpose, to gain statistical power, we extended the postnatal period to 12 months.
Statistical analysis.
To estimate the relative importance of genetic and environmental effects, we considered up to four female siblings simultaneously in family clusters. The family clusters included siblings who shared at least one parent, allowing full and half-siblings within a family. No individual was included in more than one cluster, and no known sibling relations existed between clusters. If half-siblings were clustered into more than one family, only the largest family was included to avoid duplicate entries. If a family cluster consisted of more than four individuals, four individuals were randomly selected.
As in the classical twin design, monozygotic and dizygotic twins were assumed to share 100% and 50%, respectively, of their additive genetic factors (A); the corresponding values were 50% for full siblings, and 25% for maternal and paternal half-siblings. The shared environment (C) was modeled to be fully shared by all sibling types except paternal half-siblings, where it was assumed to be unshared, as Swedish half-siblings are much more likely to live with their mother (33). Individual specific environment (E) was modeled to be unique to each individual. Relying on these assumptions, we could determine the expected correlation structures for each specific type of family cluster, depending on the sibling types that were included. We again used the liability-threshold approach for analysis of the binary traits. We fitted univariate models in which the variance in each disease, separately, was modeled to be due to A, C, and E. We then fitted bivariate models in which the variance and covariance in each trait were simultaneously modeled to be due to A, C, and E. To estimate how much of the variance in one disorder could be attributed to A, C, and E in common with the other disorder and A, C, and E unique to perinatal depression, we used a Cholesky decomposition approach (25). As this was modeled in a regression framework, we adjusted the prevalences for whether the family included half-siblings as well as for birth year (both linear and squared). Nonperinatal depression was additionally adjusted for time at risk (linear and squared) and perinatal depression for number of offspring.
Data were prepared using SAS, version 9.3 (SAS Institute, Cary, N.C.). Analyses were performed using the OpenMx package (openmx.psyc.virginia.edu) in R 3.0.2 (cran.r-project.org). In the twin-only analysis, missing values were handled by full information maximum likelihood. No missing values existed for the diseases and covariates in the sibling analyses.
The study was approved by the regional ethics committee in Stockholm.
Results
Classical Twin Design
The observed occurrence of perinatal depression in the twin sample (a total score ≥12 on the lifetime version of the Edinburgh Postnatal Depression Scale) was 7.6% (Table 1). In the ACE model, there was no significant common environmental effect, and therefore we fitted a model where the C parameter was fixed at zero (an AE model), which did not fit the data significantly worse (χ2=0.00, df=1, p=1.00). The heritability of perinatal depression based on the lifetime version of the Edinburgh Postnatal Depression Scale was estimated at 54% (95% CI=35%–70%), with the remaining variance due to nonshared environment (46%) (Table 1).
| Estimated Variance | Tetrachoric Correlation | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| Additive Genetic (A) | Shared Environment (C) | Nonshared Environment (E) | Monozygotic Twins | Dizygotic Twins | |||||||
| Model | Outcome Occurrence | Estimated Variance | 95% CIb | Estimated Variance | 95% CIb | Estimated Variance | 95% CIb | rt | SE | rt | SE |
| ACE | 7.6% | 0.54 | 0.00–0.70 | 0.00 | 0.00–0.46 | 0.46 | 0.30–0.66 | 0.55 | 0.09 | 0.22 | 0.14 |
| AEc | 7.6% | 0.54 | 0.35–0.70 | NA | 0.46 | 0.31–0.65 | 0.55 | 0.09 | 0.22 | 0.14 | |
a
Perinatal depression was defined as a score ≥12 on the lifetime version of the Edinburgh Postnatal Depression Scale.
b
Profile likelihood confidence intervals.
c
An AE model in which the C parameter was fixed at zero was considered because the estimate for the C component was not significant in the initial ACE model.
Sibling Design
Based on treatment contacts, the observed occurrence of perinatal depression was 0.6% (Table 2). The heritability of perinatal depression was estimated at 44% (95% CI=35%−52%), with the remaining variance attributable to nonshared environment. The heritability of nonperinatal depression was estimated at 32% (95% CI=24%−41%), with the remaining variance attributable to shared environment (6%) and nonshared environment (62%).
| Estimated Variance (95% CI) | |||||
|---|---|---|---|---|---|
| Disorder | Model | Outcome Occurrence | Additive Genetic (A) | Shared Environment (C) | Nonshared Environment (E) |
| Perinatal depression | ACE | 0.6% | 0.44 (0.35–0.52) | 0.00 (0.00–0.01) | 0.56 (0.48–0.64) |
| Tetrachoric correlations | |||||
| Rho | SE | Positive concurrent pairsb | |||
| Monozygotic twins | 0.45 | 0.15 | 2 | ||
| Dizygotic twinsc | –0.83 | 0.18 | 0 | ||
| Full siblings | 0.23 | 0.03 | 56 | ||
| Maternal half-siblings | 0.01 | 0.07 | 4 | ||
| Paternal half-siblings | 0.05 | 0.07 | 5 | ||
| Perinatal depression | AEd | 0.6% | 0.44 (0.35–0.52) | NA | 0.56 (0.47–0.64) |
| Tetrachoric correlations | |||||
| Rho | SE | Positive concurrent pairsb | |||
| Monozygotic twins | 0.45 | 0.15 | 2 | ||
| Dizygotic twinsc | –0.83 | 0.18 | 0 | ||
| Full siblings | 0.23 | 0.03 | 56 | ||
| Maternal half-siblings | 0.01 | 0.07 | 4 | ||
| Paternal half-siblings | 0.05 | 0.07 | 5 | ||
| Nonperinatal depression | ACE | 5.4% | 0.32 (0.24–0.41) | 0.06 (0.02–0.10) | 0.62 (0.57–0.66) |
| Tetrachoric correlations | |||||
| Rho | SE | Positive concordant pairsb | |||
| Monozygotic twins | 0.52 | 0.06 | 31 | ||
| Dizygotic twins | 0.15 | 0.10 | 9 | ||
| Full siblings | 0.22 | 0.01 | 1,834 | ||
| Maternal half-siblings | 0.13 | 0.02 | 296 | ||
| Paternal half-siblings | 0.05 | 0.02 | 239 | ||
a
Perinatal depression and nonperinatal depression were defined on the basis of register-based data on hospital admissions for depression (see the Method section). Confidence intervals are Wald type, and standard errors are calculated using the delta method.
b
Number of pairs consisting of two individuals positive for the disorder.
c
No concordant dizygotic twin pairs for perinatal depression.
d
An AE model in which the C parameter was fixed at zero was considered because the estimate for the C component was not significant in the initial ACE model.
In a bivariate model all C parameters, except C unique for nonperinatal depression, was estimated close to zero. Therefore, we fitted an AE model in which these parameters were set to zero, and the model fit did not deteriorate (χ2=0.53, df=2, p=0.77). The bivariate analysis revealed that 14% of the total variance (or 33% of the genetic variance) in perinatal depression was unique for perinatal depression and not in common with nonperinatal depression (Table 3 and Figure 1).
| Estimated Variance (95% CI) | |||||||
|---|---|---|---|---|---|---|---|
| Disorder | Model | Additive Genetic (A), Uniqueb | Additive Genetic (A), in Commonc | Environment | |||
| Shared (C), Uniqueb | Shared (C), in Commonc | Nonshared (E), Uniqueb | Nonshared (E), in Commonc | ||||
| Perinatal depression | ACE | 0.16 (0.13–0.18) | 0.26 (0.20–0.32) | 0.00 (0.00–0.00) | 0.00 (0.00–0.01) | 0.42 (0.38–0.45) | 0.17 (0.14–0.20) |
| Perinatal depression | AEd | 0.14 (0.12–0.16) | 0.28 (0.22–0.35) | NA | NA | 0.42 (0.38–0.45) | 0.16 (0.13–0.19) |
| Nonperinatal depression | ACEe | 0.11 (0.09–0.14) | 0.24 (0.20–0.28) | 0.04 (0.02–0.07) | NA | 0.43 (0.41–0.45) | 0.17 (0.14–0.20) |
| Tetrachoric correlations | |||||||
| Rho | SE | Positive concordant pairsf | |||||
| Monozygotic twins | 0.33 | 0.10 | 9 | ||||
| Dizygotic twins | 0.05 | 0.29 | 1 | ||||
| Full siblings | 0.16 | 0.01 | 381 | ||||
| Maternal half-siblings | 0.08 | 0.03 | 75 | ||||
| Paternal half-siblings | 0.06 | 0.02 | 67 | ||||
a
Variance explained in either the investigated disorder by factors unique for the disorder, or shared with the other disorder. Confidence intervals are Wald type, and standard errors are calculated using the delta method.
b
Variance explained in the disorder by component unique for that disorder and not in common with the other disorder.
c
Variance explained in the disorder by component in common with the other disorder.
d
An AE model in which the C parameter was fixed at zero was considered because the estimate for the C component was not significant in the initial ACE model (except C unique for nonperinatal depression).
e
An ACE model in which the effects of the C parameter involving perinatal depression (C in common) was fixed at zero.
f
Number of pairs consisting of one individual positive for perinatal depression and the other positive for nonperinatal depression. Each pair contributes with two combinations; perinatal depression sibling 1 versus nonperinatal depression sibling 2, and vice versa.
FIGURE 1. Estimated Variance in Perinatal Depressiona
a Variance accounted for by genetic, shared environmental, and nonshared environmental effects for perinatal depression, either unique for perinatal depression or in common with nonperinatal depression. No variance accounted for by shared environmental effects.
The heritability of antenatal depression was estimated at 37% (95% CI=27%−47%), with the remaining variance attributable to nonshared environment, while the heritability of postnatal depression was estimated at 40% (95% CI=31%−49%), with the remaining variance attributable to nonshared environment (Table 4).
| Estimated Variance (95% CI) | |||||
|---|---|---|---|---|---|
| Disorder | Model | Outcome Occurrence | Additive Genetic (A) | Shared Environment (C) | Nonshared Environment (E) |
| Antenatal depression | ACE | 0.3% | 0.36 (0.27–0.45) | 0.02 (–0.04–0.09) | 0.62 (0.54–0.70) |
| Antenatal depression | AEb | 0.3% | 0.37 (0.27–0.47) | NA | 0.62 (0.51–0.73) |
| Tetrachoric correlations | |||||
| Rho | SE | Positive concordant pairsd | |||
| Monozygotic twins | –0.82 | NAc | 0 | ||
| Dizygotic twins | –0.64 | NAc | 0 | ||
| Full siblings | 0.24 | 0.00 | 18 | ||
| Maternal half-siblings | 0.15 | 0.17 | 2 | ||
| Paternal half-siblings | 0.04 | 0.11 | 2 | ||
| Postnatal depression | ACE | 0.4% | 0.40 (0.27–0.52) | 0.00 (–0.03–0.03) | 0.60 (0.49–0.71) |
| Postnatal depression | AEb | 0.4% | 0.40 (0.31–0.49) | NA | 0.60 (0.51–0.69) |
| Tetrachoric correlations | |||||
| Rho | SE | Positive concordant pairsd | |||
| Monozygotic twins | 0.64 | 0.13 | 3 | ||
| Dizygotic twins | –0.84 | NAc | 0 | ||
| Full siblings | 0.19 | 0.00 | 33 | ||
| Maternal half-siblings | 0.08 | 0.07 | 6 | ||
| Paternal half-siblings | 0.07 | 0.09 | 3 | ||
a
Antenatal depression (during pregnancy) and postnatal depression (within 12 months postpartum) were defined according to register-based data on hospital admissions for depression (see the Method section). Confidence intervals are Wald type, and standard errors are calculated using the delta method.
b
An AE model in which the C parameter was fixed at zero was considered because the estimate for C was zero in the ACE model.
c
No standard errors were obtained because there were no positive concordant pairs.
d
Number of pairs consisting of two individuals positive for the disorder.
Discussion
To our knowledge, this is the largest and most comprehensive genetic epidemiological study of perinatal depression yet reported. Using Swedish national cohorts, we estimated the heritability of perinatal depression with two different approaches. Our classical twin design estimated the univariate heritability of perinatal depression at 54%, and the sibling study estimated it at 44%. Despite the marked difference in observed occurrence of perinatal depression in the two designs (0.6% and 7.6%), the heritability estimates are similar and the confidence intervals overlap, which is consistent with both approaches capturing the same underlying liability for perinatal depression. The different observed occurrence rates are likely explained by the different methodologies—self-report in the twin design and register-based treatment contacts in the sibling design. Thus, the sibling design did not account for women who did not seek treatment for perinatal depression but who would endorse symptoms on a self-report questionnaire (34).
Division of the perinatal period allowed for estimation of the heritability of antenatal depression at 37% (95% CI=27%−47%) and postnatal depression at 40% (95% CI=31%−49%), with the remaining variance explained by nonshared environment in both disorders (Table 4). Thus the variance in both antenatal and postnatal depression displayed a pattern similar to the variance in perinatal depression as a whole.
To our knowledge, there has only been one previous heritability study of perinatal depression (15). This Australian twin study (N=1,676) estimated the heritability of lifetime postnatal depression at 25% (95% CI=13%−42%). Our estimates for lifetime perinatal depression at 54% (95% CI=35%−70%) in the twin design and at 44% (95% CI=35%−52%) in the sibling design indicate a larger genetic contribution than in the Australian study.
We estimated the heritability of nonperinatal depression at 32%, which is slightly lower than previous estimates of major depressive disorder (12). However, we included only parous women, and we separated the perinatal and nonperinatal depressive episodes.
In our bivariate heritability analysis of perinatal depression and nonperinatal depression, 14% of the variance in perinatal depression was explained by genetic factors unique for perinatal depression and 28% by genetic factors shared with nonperinatal depression. In other words, of the total genetic variation for perinatal depression, two-thirds is shared with nonperinatal depression and one-third is unique for perinatal depression.
The heritability estimates of perinatal depression in these two analyses may have particularly important implications. A critical issue facing genetic research in unipolar mood disorders is etiological heterogeneity. It is likely that there are multiple “types” of persistent depressive disorders. Considering these disorders as a single entity may effectively combine different sets of genetic and environmental etiological processes, resulting in higher prevalence and lower heritability (35–37). This combination is arguably unfavorable for genomic discovery efforts (38).
Thus, we hypothesize that perinatal depression represents a form of unipolar mood disorder that could be prioritized for genomic discovery efforts. In effect, we suggest a “divide and conquer” approach to understanding the genomics of unipolar mood disorders. Appropriately powered studies of perinatal depression could deliver genomic findings that are important to disentangling its etiology and that have potential relevance to major depressive disorder. We note that women with perinatal depression are readily ascertained clinically, enabling efficient accrual of large samples. Ultimately, improved identification of women at risk for perinatal depression could lead to targeted interventions to prevent, identify, and more effectively treat perinatal depression in order to minimize adverse sequelae for mother and child.
This study has several strengths. It used both classical twin and sibling designs, and different tools to measure perinatal depression: the validated Edinburgh Postnatal Depression Scale in 3,427 female twins, and treatment contacts from national Swedish registers in over 580,000 sisters, which allowed for separation of depressive illness into perinatal depression and nonperinatal depression, and further division into antenatal and postnatal depression. Additionally, sensitivity analyses suggested that the unique genetic component seen in perinatal depression was not explained by bipolar disorder or schizophrenia (see appendix 1 in the data supplement that accompanies the online edition of this article) and that the unique genetic component was linked to the actual pregnancy (see appendix 2 in the data supplement ).
This study also has limitations. The Swedish National Patient Register does not include outpatient admissions before 2001 and it includes no primary care data. The percentage of women being treated for depression exclusively in primary care in Sweden is not known, but the observed occurrence of perinatal depression based on treatment contacts is likely an underestimate of the true prevalence. Furthermore, depression identified using the National Patient Register will likely be on the more severe end of the spectrum compared with depression identified using the Edinburgh Postnatal Depression Scale. If the assumption of an underlying continuous liability in the threshold model is true, this should not affect the heritability estimates. Indeed, increasing or decreasing the Edinburgh Postnatal Depression Scale cutoff to capture more or less severe depressive illness did not change the heritability estimates (see appendix 3 in the data supplement ). The observed outcome occurrence of perinatal depression among the twins using the Edinburgh Postnatal Depression Scale was 7.6%, which is lower than observed in other studies (10%−15%) (1–4). This could be due to the participants being exclusively twins and to use of retrospective assessments rather maternal health care visit assessments. When estimating the heritability of antenatal and postnatal depression, respectively, the postnatal period had to be expanded to 12 months to include enough cases to allow analysis. While this deviates from the traditional definition of perinatal depression, which covers a 6-month postnatal period, analysis of perinatal depression using a 12-month postnatal period is consistent with a broader definition and revealed almost identical results, with the heritability estimated at 43% (95% CI=34%−51%) and the remaining variance explained by nonshared environment (57%; 95% CI=49%−66%). We were also not able to restrict the postnatal period to the first 4 or 6 weeks. However, early onset of symptoms may not lead to contact with the health care system within this period unless symptoms persist and could therefore remain undetected when using treatment contact data. Examination of differences regarding timing of onset of symptoms in pregnancy versus the postpartum period has been a central issue in recent work (39), and future research should focus on further elucidating the genetic and biological contributions to the timing of onset of symptoms.
These findings provide important information that will assist clinicians as they counsel their patients regarding the risk and prognosis of perinatal mood disorders. For example, the heritability of a disorder has a direct translational impact in discussions between clinicians and patients. Most patients will ask fundamental questions, such as “Why do I have perinatal depression?,” “Was it my fault?,” and “What is the risk next time?” This study highlights the critical need for clinicians providing obstetrical care to obtain detailed information regarding the patient’s personal and family history of any psychiatric illness that began in a perinatal period. Integration of genetic risk with environmental influences is vital for the appropriate tailoring of individual treatment and discussions of prognosis.
In conclusion, we report the largest heritability studies of perinatal depression to date, including the first bivariate heritability study of perinatal and nonperinatal depression, revealing that one-third of the genetic variance is unique for perinatal depression. We believe that perinatal depression represents a form of unipolar mood disorder that can be utilized by clinicians in discussions with their patients and could be prioritized for genomic discovery efforts.
Supplementary Material
File (appi.ajp.2015.15010085.ds001.pdf)
- Download
- 114.99 KB
References
1.
Flynn HA, Davis M, Marcus SM, et al: Rates of maternal depression in pediatric emergency department and relationship to child service utilization. Gen Hosp Psychiatry 2004; 26:316–322
2.
Gavin NI, Gaynes BN, Lohr KN, et al: Perinatal depression: a systematic review of prevalence and incidence. Obstet Gynecol 2005; 106:1071–1083
3.
Marmorstein NR, Malone SM, Iacono WG: Psychiatric disorders among offspring of depressed mothers: associations with paternal psychopathology. Am J Psychiatry 2004; 161:1588–1594
4.
Meltzer-Brody S, Stuebe A: The long-term psychiatric and medical prognosis of perinatal mental illness. Best Pract Res Clin Obstet Gynaecol 2014; 28:49–60
5.
Bernstein IH, Rush AJ, Yonkers K, et al: Symptom features of postpartum depression: are they distinct? Depress Anxiety 2008; 25:20–26
6.
Wisner KL, Moses-Kolko EL, Sit DK: Postpartum depression: a disorder in search of a definition. Arch Women Ment Health 2010; 13:37–40
7.
Lecrubier Y: The burden of depression and anxiety in general medicine. J Clin Psychiatry 2001; 62(suppl 8):4–9
8.
Murray CJ, Lopez AD: Alternative projections of mortality and disability by cause, 1990–2020: Global Burden of Disease Study. Lancet 1997; 349:1498–1504
9.
Murray CJ, Lopez AD: Evidence-based health policy: lessons from the Global Burden of Disease Study. Science 1996; 274:740–743
10.
Ekman M, Granström O, Omérov S, et al: The societal cost of depression: evidence from 10,000 Swedish patients in psychiatric care. J Affect Disord 2013; 150:790–797
11.
Wray NR, Gottesman II: Using summary data from the Danish national registers to estimate heritabilities for schizophrenia, bipolar disorder, and major depressive disorder. Front Genet 2012; 3:118
12.
Sullivan PF, Neale MC, Kendler KS: Genetic epidemiology of major depression: review and meta-analysis. Am J Psychiatry 2000; 157:1552–1562
13.
Levinson DF, Mostafavi S, Milaneschi Y, et al: Genetic studies of major depressive disorder: why are there no genome-wide association study findings and what can we do about it? Biol Psychiatry 2014; 76:510–512
14.
Major Depressive Disorder Working Group of the Psychiatric GWAS Consortium: A mega-analysis of genome-wide association studies for major depressive disorder. Mol Psychiatry 2013; 18:497–511
15.
Treloar SA, Martin NG, Bucholz KK, et al: Genetic influences on post-natal depressive symptoms: findings from an Australian twin sample. Psychol Med 1999; 29:645–654
16.
Murphy-Eberenz K, Zandi PP, March D, et al: Is perinatal depression familial? J Affect Disord 2006; 90:49–55
17.
Forty L, Jones L, Macgregor S, et al: Familiality of postpartum depression in unipolar disorder: results of a family study. Am J Psychiatry 2006; 163:1549–1553
18.
Bloch M, Rubinow DR, Schmidt PJ, et al: Cortisol response to ovine corticotropin-releasing hormone in a model of pregnancy and parturition in euthymic women with and without a history of postpartum depression. J Clin Endocrinol Metab 2005; 90:695–699
19.
Bloch M, Daly RC, Rubinow DR: Endocrine factors in the etiology of postpartum depression. Compr Psychiatry 2003; 44:234–246
20.
Lichtenstein P, De Faire U, Floderus B, et al: The Swedish Twin Registry: a unique resource for clinical, epidemiological, and genetic studies. J Intern Med 2002; 252:184–205
21.
Magnusson PK, Almqvist C, Rahman I, et al: The Swedish Twin Registry: establishment of a biobank and other recent developments. Twin Res Hum Genet 2013; 16:317–329
22.
Meltzer-Brody S, Boschloo L, Jones I, et al: The EPDS-Lifetime: assessment of lifetime prevalence and risk factors for perinatal depression in a large cohort of depressed women. Arch Women Ment Health 2013; 16:465–473
23.
Cox JL, Holden JM, Sagovsky R: Detection of postnatal depression: development of the 10-item Edinburgh Postnatal Depression Scale. Br J Psychiatry 1987; 150:782–786
24.
Wisner KL, Parry BL, Piontek CM: Clinical practice: postpartum depression. N Engl J Med 2002; 347:194–199
25.
Neale MC, Cardon LR, North Atlantic Treaty Organization, Scientific Affairs Division: Methodology for Genetic Studies of Twins and Families. Dordrecht, the Netherlands, Kluwer Academic Publishers, 1992
26.
Divgi DR: Calculation of the tetrachoric correlation coefficient. Psychometrika 1979; 44:169–172
27.
Kendler KS, Neale MC, Kessler RC, et al: A test of the equal-environment assumption in twin studies of psychiatric illness. Behav Genet 1993; 23:21–27
28.
Cnattingius S, Ericson A, Gunnarskog J, et al: A quality study of a medical birth registry. Scand J Soc Med 1990; 18:143–148
29.
Ekbom A: The Swedish Multi-Generation Register. Methods Mol Biol 2011; 675:215–220
30.
Ludvigsson JF, Andersson E, Ekbom A, et al: External review and validation of the Swedish national inpatient register. BMC Public Health 2011; 11:450
31.
Sharma V, Sharma P: Postpartum depression: diagnostic and treatment issues. J Obstet Gynaecol Can 2012; 34:436–442
32.
Beck CT: Theoretical perspectives of postpartum depression and their treatment implications. MCN Am J Matern Child Nurs 2002; 27:282–287
33.
Holmberg M: Fakta om den svenska familjen: sammansättning och förändringar från barndom till ålderdom. Örebro, Sweden, Statistiska centralbyrån (SCB), 1994
34.
Leung SS, Leung C, Lam TH, et al: Outcome of a postnatal depression screening programme using the Edinburgh Postnatal Depression Scale: a randomized controlled trial. J Public Health (Oxf) 2011; 33:292–301
35.
Lewis CM, Ng MY, Butler AW, et al: Genome-wide association study of major recurrent depression in the UK population. Am J Psychiatry 2010; 167:949–957
36.
Sullivan PF: The psychiatric GWAS consortium: big science comes to psychiatry. Neuron 2010; 68:182–186
37.
Wray NR, Pergadia ML, Blackwood DH, et al: Genome-wide association study of major depressive disorder: new results, meta-analysis, and lessons learned. Mol Psychiatry 2012; 17:36–48
38.
Sullivan PF, Daly MJ, O’Donovan M: Genetic architectures of psychiatric disorders: the emerging picture and its implications. Nat Rev Genet 2012; 13:537–551
39.
Postpartum Depression: Action Towards Causes and Treatment (PACT) Consortium: Heterogeneity of postpartum depression: a latent class analysis. Lancet Psychiatry 2015; 2:59–67
Information & Authors
Information
Published In
History
Received: 21 January 2015
Revision received: 1 June 2015
Accepted: 15 June 2015
Published online: 4 September 2015
Published in print: February 01, 2016
Authors
Competing Interests
Dr. Meltzer-Brody has received research support from Sage Therapeutics. Dr. Landén has received compensation for lectures from AstraZeneca, Bayer, Biophausia, Bristol Myers-Squibb, Lundbeck, Eli Lilly, Wyeth, and Servier and has served as an advisory board member for AstraZeneca and Lundbeck. The other authors report no financial relationships with commercial interests.
Funding Information
The Swedish Twin Registry is financially supported by Karolinska Institutet. The present study was supported by grants from the Swedish Medical Research Council (K2014-62X-14647-12-51 and K2010-61P-21568-01-4), the Swedish Foundation for Strategic Research, the Swedish Brain Foundation, and NIMH (K23 MH085165).
Metrics & Citations
Metrics
Citations
Export Citations
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.
For more information or tips please see 'Downloading to a citation manager' in the Help menu.
There are no citations for this item
View Options
View options
PDF/ePub
View PDF/ePubGet Access
Login options
Already a subscriber? Access your subscription through your login credentials or your institution for full access to this article.
Personal login Institutional Login Open Athens loginNot a subscriber?
PsychiatryOnline subscription options offer access to the DSM-5-TR® library, books, journals, CME, and patient resources. This all-in-one virtual library provides psychiatrists and mental health professionals with key resources for diagnosis, treatment, research, and professional development.
Need more help? PsychiatryOnline Customer Service may be reached by emailing [email protected] or by calling 800-368-5777 (in the U.S.) or 703-907-7322 (outside the U.S.).