Affective Temperaments May Predict Postpartum Depression: A Preliminary Study

Erson AKSU; Elmas BEYAZYÜZ

doi:10.4274/nkmj.galenos.2020.782360

ABSTRACT

Aim:

Postpartum depression (PPD) is a serious clinical condition, which affects both mother and baby as well as partners. Although PPD is accepted as a mood disorder, there have been limited numbers of studies that investigated affective temperaments in PPD and pregnancy. Moreover, no study has been carried out on the subject of whether affective temperaments assessed during pregnancy can predict PPD.

Materials and Methods:

One hundred-fourteen pregnant women were included in present study considering specific criteria. Patients were assessed with the Temperament Evaluation of Memphis, Pisa, Paris and San Diego auto-questionnaire and socio-demographic forms during the third trimester of pregnancy. In the following four weeks of delivery, patients were assessed with the Edinburgh Postpartum Depression Scale for evaluating the presence of tendency to PPD.

Results:

The patients who had a tendency to PPD had significantly higher scores on all affective temperament scores except hyperthymic temperament. In logistic regression analysis, it was found that higher scores of cyclothymic temperament and anxious temperament (AT) were associated with tendency to PPD (Odds ratio: 1.26 and 1.47).

Conclusion:

Cyclothymic and anxious temperaments are candidates for predicting tendency to PPD. Pregnant women, who have higher scores specifically for cyclothymic and ATs should be considered in terms of affective disorders.

Keywords:

Postpartum, depression, affective, temperament

INTRODUCTION

Depressive disorder is one of the most common cause of disability in women in the whole World and the prevalence of depression is twice the number in women compared to men¹. Postpartum phase is also an important time interval for developing depression among women. The Diagnostic and Statistical Manual of Mental Disorders-5^th Edition (APA 20013) describes the postpartum depression (PPD) as a depressive state that has symptoms ranging from moderate to severe, which begins during four weeks after delivery^2-4.

Postpartum psychiatric disorders consist of PPD, postpartum blues and postpartum psychosis⁵.PPD causes severe social and occupational functional impairments and is closely associated with problems in interactions with partner, family and baby^6-8. The strongest risk factors for PPD have been identified as antenatal depression or anxiety, history of major depressive disorder, problems in marriage, poor social support and stressing life events^9,10. Furthermore patients with a history of PPD have been reported to have higher percentages of PPD subsequent to deliveries^11,12.

Currently, affective temperaments are considered to be predictors of some mood disorders¹³. Depressive temperament (DT) was reported to be related to depressive disorder, and irritable temperament (IT), hyperthymic temperament (HT) were considered to be associated with bipolar 1 disorder, and cyclothymic temperament (CT) was regarded as a tendency to bipolar 2 disorder^13,14. However, there have been limited numbers of studies that investigated affective temperaments during pregnancy and postpartum period. Yazici et al.¹⁵reported that affective temperament scores of pregnant women were different from those of healthy control group. Another current study reported that pregnancy and postpartum periods correlated with HT in women without psychiatric diagnosis¹⁵. There has been only one study that assessed affective temperaments in patients with PPD¹⁶. However, no study assessing whether affective temperaments evaluated during pregnancy may predict PPD has been carried out.

In the present study, we aimed to investigate whether there would be associations between affective temperaments assessed during pregnancy and PPD. We hypothesized that some affective temperaments could predict PPD.

GİRİŞ

Depresif bozukluk, tüm dünyada kadınlarda en yaygın güçsüzlük nedenlerinden biridir ve depresyon prevalansı kadınlarda erkeklere göre iki kat fazladır¹. Postpartum dönem aynı zamanda kadınlar arasında depresyonun gelişmesi için önemli bir zaman aralığıdır. Mental Bozuklukların Tanısal ve Sayımsal El Kitabı-5. baskısı (APA 20013), postpartum depresyonu (PPD), doğumdan sonraki dört haftada başlayan, orta ila şiddetli arasında değişen semptomları olan depresif bir durum olarak tanımlar^2-4.

Postpartum psikiyatrik bozukluklar, PPD, postpartum hüzün ve postpartum psikozdan oluşur⁵. PPD, ciddi sosyal ve mesleki işlevsel bozukluklara neden olur ve partner, aile ve bebekle etkileşimdeki problemlerle yakından ilişkilidir^6-8. PPD için en güçlü risk faktörleri antenatal depresyon veya anksiyete, majör depresif bozukluk öyküsü, evlilik sorunları, zayıf sosyal destek ve stresli yaşam olayları olarak belirlenmiştir^9,10. Ayrıca, PPD öyküsü olan hastaların doğumları takiben daha yüksek PPD yüzdelerine sahip oldukları bildirilmiştir^11,12.

Şu anda, afektif mizaçların bazı duygudurum bozukluklarının yordayıcıları olduğu düşünülmektedir¹³. Depresif mizacın (DT) depresif bozuklukla ilişkili olduğu, irritabl mizaç (IT), hipertimik mizacın (HT) bipolar 1 bozukluk ile ilişkili olduğu ve siklotimik mizacın (CT) bipolar 2 bozukluğuna eğilim olarak kabul edildiği bildirilmiştir^13,14. Bununla birlikte, gebelikte ve doğum sonrası dönemde afektif mizaçları araştıran sınırlı sayıda çalışma vardır. Yazici ve ark.¹⁵ gebelerin afektif mizaç skorlarının sağlıklı kontrol grubunun skorlarından farklı olduğunu bildirmişlerdir. Bir başka çalışmada, psikiyatrik tanısı olmayan kadınlarda gebelik ve postpartum dönemlerin HT ile ilişkili olduğu bildirilmiştir¹⁵. PPD hastalarında afektif mizaçları değerlendiren sadece bir çalışma yapılmıştır¹⁶. Bununla birlikte, gebelik sırasında değerlendirilen afektif mizaçların PPD’yi öngörüp öngöremeyeceğini değerlendiren bir çalışma yapılmamıştır.

Bu çalışmada, gebelikte değerlendirilen afektif mizaçlar ile PPD arasında ilişki olup olmayacağını araştırmayı amaçladık. Bazı afektif mizaçların PPD’yi öngörebileceğini varsaydık.

MATERIALS AND METHODS

The present study was conducted at Tekirdağ Namık Kemal University Faculty of Medicine, Departments of Gynecology and Obstetrics and Psychiatry. It was approved by Tekirdağ Namık Kemal University Non-Invasive Clinic Research Ethical Committee (date and approval number: 2018/124/08/15).

Inclusion criteria were defined as being at the third period of pregnancy, having no previous or current psychiatric diagnosis, having enough education for being able to complete the self-assessment tests that were used in the present study, and being willing to be participate in the study. Patients who had previous or current psychiatric diagnosis, who had insufficient knowledge for understanding the aim of the study as well as tests that were used in the study, and who were unwilling to participate in the study were excluded. According to inclusion and exclusion criteria, 170 female patients, who were at the 3^rd trimester of their pregnancy, were initially involved in the present study. All of the patients completed the Temperament Evaluation of Memphis, Pisa, Paris and San Diego Auto-questionnaire (TEMPS-A) form and socio-demographic form during pregnancy. After delivery, all patients, who completed TEMPS-A, were evaluated by the Edinburgh Postpartum Depression Scale (EPDS) at the routine controls in the four weeks after delivery. As 114 patients completed the second round, 114 patients were finally included in the study. All patients signed written informed consent form before participating in the study.

Assessment Tools

Socio-demographic Form

This form was created by authors in the light of the literature. This form consists of the data on age, duration of marriage, having a child, education and occupation status, family structure, income rate, place of birth, status of cigarette, and alcohol use.

Temperament evaluation of Memphis, Pisa, Paris and San Diego Auto-questionnaire

The TEMPS-A is a scale that was originally designed by Vahip et al.¹⁷ and then adopted into Turkish by Vahip et al.¹⁷. In the present study, this scale was used to evaluate the scores of subdimensions. It is a self-assessment scale, involving “true” or “false” indications that ask about mood and temperament properties of the entire life of the individual. The subdimensions consist of depressive, cyclothymic, hyperthymic, irritable (DT, CT, HT, IT), and anxious temperaments (AT).

The Edinburgh Postpartum Depression Scale

EPDS was created by Cox et al., and translated and adapted to Turkish by Engindeniz et al.¹⁸ (1997). The purpose of the scale is to evaluate women’s PPD levels and it consists of ten items rated on a four-point Likert scale ranging from 0 to 3. The lowest total score that can be obtained is 0, and the highest is 30. Individuals, who score 13 or more, are considered to be at risk of depression.

Statistical Analysis

Statistical analyses were conducted with R 3.5.3, SPSS 23.0., STATA 14.0 and G * Power 3.1. Power analysis was performed to determine the power of the study. We defined a new variable by using EPSD score with its cut-off point. According to that, we tried to predict the redefined EPSD variable by using the Binary Outcome Logistic Regression. This new variable was also used to determine whether the mean of EPDS dimensions differed. For the assumptions of parametric tests, the Kolmogorov-Smirnov test was used to test for data normality and the box plot was also used. The Levine test was used for the variance homogeneity.

Power Analysis

The Mann-Whitney U test statistic with allocation ratio 0.29 was used as a test statistic. A sample size of 20 achieves 70% power to detect an effect size (d) of 1.3978 using the Mann-Whitney U test with a significance level (alpha) of 0.05. When the sample size is n=45, the power of the test has already achieved 95%. However, the sample size of the study was determined as 114. For the given parameters, for an alpha of 0.05 and a sample size of 114 observations, the type 2 error is 0.0003 and the power is ≌ 1.0 (Figure 1).

RESULTS

One hundred fourteen patients were included in the present study. The mean age of participants was 29.72±5.73 years. The median scores of DT, CT, HT, IT and AT were 4, 6, 10, 2 and 7, respectively. The numbers of patients, who scored 13 and higher on EPDS, were 26 (22.8%). The data of demographical and clinical variables of patients were demonstrated in Table 1.

The Kolmogorov-Smirnov test was used to examine the normality of the presence-absence of tendency to PPD distributions. None of them distributed normally. Both the Kolmogorov-Smirnov test and the box-plots showed that presence- absence of tendency to PPD distributions were skewed (Figure 2). For two independent samples comparison, the Mann-Whitney U test was applied when the assumption of normal distribution did not fit. According to the presence- absence of tendency to PPD based on EPDS, there were statistically significant differences between DT scores (absence: 4.34±2.83, presence: 6.35±2.91), CT scores (absence: 6.64±5.16, presence: 8.96±4.86), IT scores (absence: 2.80±2.85, presence: 5.27±4.31) and AT scores (absence: 6.12±6.00, presence: 14.35±5.76) (Table 2).

The general form of the logistic regression model is:

logit(P)=a+b1x1+b2x2+b3x3+...+bkxk

where logit(P)=Y is a dependent variable and x1,x2,x3,...,xk are independent variables.

The parameters of b1,b2,b3,...,bk are the logistic regression coefficients. The binary response variable is taken as 0 or 1. The value 1 means the presence of PPD and value 0 indicates the absence of PPD. The P is denoted as the probability of presences of tendency to PPD. A pseudo R² greater than 2 indicates a relatively good fit; equals to 1 indicates a perfect fit and equal to 0 means no relationship. According to the Table 3, the model has a relatively good fit (pseudo R²=0.48). The Cox and Snell’s R²and Nagelkerke R²indicates that the independent variables can explain the dependent variables with the values of approximately 40% and 61%. For all that, the predicted accuracy was 94% for the absence of depression, the predicted accuracy was 73% for the presence of tendency to PPD and the overall predicted accuracy was 89%. The model was significant at 5% significance level. CT, AT, having a child, and place of birth had a statistically significant effect on the presence of tendency to PPD (Table 3).

DISCUSSION

In the present study, our main findings were as follows; except HT, all subdimensions of TEMPS-A were significantly higher in the patient group who scored ≥13 at the EPDS, which reflects tendency to PPD. Moreover, we found that higher scores of CT and AT were associated with tendency to PPD.

PPD is a severe mental disorder which affects both mother and child^19-21. The exact etiology of PPD is considered unclear; however, biological factors such as ovarian steroids, oxytocin and glucocorticoids as well as other neurotransmitters, which are also associated with mood disorders, have been reported to be etiological factors⁶.Affective temperaments are among well-established factors for the development of affective disorders^22,23. Akiskal ve Akiskal²⁴created the modern concept of affective temperaments for identifying all spectrum of affective situations from healthy reactions to major affective disorders.

There have been two studies that investigated affective temperaments in pregnant women. In the first study, Yazici et al.¹⁵compared affective temperaments between pregnant women and healthy women and reported that cyclothymic, irritable and AT scores of the pregnant women were significantly lower in pregnant women, and they concluded that differences of affective temperaments could be associated with trimesters of pregnancy.Yazici et al.¹⁶ investigated affective temperaments between pregnant women in different trimesters, women who were at prenatal period and postpartum period, and age-matched healthy women. The researchers concluded that pregnancy and postpartum periods correlated with HT characteristics in women without active psychiatric diagnosis.

In our study, we compared affective temperaments, which were obtained at third trimester between pregnant women with and without tendency to PPD, based on EPDS. We found that except HT, all subdimensions of TEMPS-A were significantly higher in the patient group who scored ≥13 at the EPDS. These temperaments were reported more frequently in affective patients with mixed episodes indicating a relationship between mixed affective episodes and simultaneous presence of inverse temperamental types²⁵.

From this perspective, we can say that affective temperaments differed in pregnant women with and without the risk for postpartum depressive disorder. However, our main purpose was to provide predictive roles of affective temperaments for developing postpartum depressive disorder. We performed logistic regression analysis and found that higher scores of CT and AT were associated with tendency to PPD (respectively Odds ratio; 1.26 and 1.47). There has been only one study that investigated affective temperaments in patients, who were diagnosed with PPD. In this study, affective temperament and the presence of postpartum depressive disorder were assessed simultaneously and it was reported that cyclothymic and AT were among the significant risk factors independently from psychosocial factors¹⁷. The methodology of our study differed from Masmoudi et al.’s²⁶ research. Firstly, we assessed affective temperaments in women at third trimester of pregnancy, who had no present or previous psychiatric disorder. Secondly, our study is a follow-up study that investigated the patients in both pregnancy and postpartum periods. Our results are first to show cyclothymic and AT as candidates for developing tendency for postpartum depressive disorder. Additionally, having a child and place of birth were found to be other risk factors for postpartum depressive disorder.

Study Limitations

Although we performed power analysis for the present study, the numbers of patients can be considered small for making a general conclusion in terms of identifying predictors for postpartum depressive disorder. This is the major limitation of our study. Including the pregnant women without psychiatric disorder, assessing the patients at the same trimester, evaluating the PPD in the first four weeks and prospective design of our study can be regarded as the strengths of our study.

TARTIŞMA

Bu çalışmada ana bulgularımız şu şekildedir; HT dışında TEMPS-A’nın tüm alt boyutları doğum sonrası depresyona yatkınlığı yansıtan EPDS’de ≥13 puan alan hasta grubunda anlamlı olarak daha yüksekti. Ayrıca, yüksek CT ve AT puanlarının PPD eğilimi ile ilişkili olduğunu bulduk.

PPD, hem anneyi hem de çocuğu etkileyen ciddi bir zihinsel bozukluktur^19-21. PPD’nin kesin etiyolojisi belirsiz kabul edilmektedir; ancak ovaryen steroidler, oksitosin ve glukokortikoidler gibi biyolojik faktörlerin yanı sıra duygudurum bozuklukları ile de ilişkili diğer nörotransmitterler etiyolojik faktörler olarak bildirilmiştir⁶. Afektif mizaçlar, afektif bozukluklarının gelişmesi için iyi bilinen faktörler arasındadır^22,23. Akiskal ve Akiskal²⁴, sağlıklı tepkilerden majör afektif bozukluklara kadar tüm afektif durumları belirlemek için modern afektif mizaç kavramını ortaya koydular.

Hamile kadınlarda afektif mizaçları inceleyen iki çalışma yapılmıştır. İlk çalışmada Yazici ve ark.¹⁵ hamile kadınlar ile sağlıklı kadınlar arasındaki afektif mizaçları karşılaştırmışlar, gebelerin siklotimik, IT ve endişeli mizaç skorlarının gebelerde anlamlı olarak daha düşük olduğunu bildirmişler ve afektif mizaç farklılıklarının gebeliğin trimesterleri ile ilişkili olabileceği sonucuna varmışlardır. Yazici ve ark.¹⁶, farklı trimesterdaki gebeler, doğum öncesi ve doğum sonrası dönemde olan kadınlar ve aynı yaştaki sağlıklı kadınlar arasındaki afektif mizaçları araştırdı. Araştırmacılar, aktif psikiyatrik tanısı olmayan kadınlarda gebelik ve postpartum dönemlerin HT özellikleriyle ilişkili olduğu sonucuna varmışlardır.

Çalışmamızda PPD eğilimi olan ve olmayan gebeler arasında üçüncü trimesterde elde edilen afektif mizaçları EPDS’ye göre karşılaştırdık. EPDS’de ≥13 puan alan hasta grubunda HT dışında TEMPS-A’nın tüm alt boyutlarının anlamlı olarak daha yüksek olduğunu bulduk. Bu mizaçlar, karışık epizodları olan afektif hastalarda daha sık bildirilmiştir, bu durum da karışık afektif epizodlar ile ters mizaç tiplerinin eşzamanlı varlığı arasında bir ilişki olduğunu göstermektedir²⁵.

Bu açıdan bakıldığında postpartum depresif bozukluk riski olan ve olmayan gebelerde afektif mizaçların farklılaştığını söyleyebiliriz. Ancak asıl amacımız postpartum depresif bozukluğun gelişmesi için afektif mizaçların yordayıcı rollerini sağlamaktı. Lojistik regresyon analizi yaptık ve yüksek CT ve kaygılı mizaç skorlarının PPD eğilimi ile ilişkili olduğunu bulduk (sırasıyla odds oranı: 1,26 ve 1,47). PPD tanısı alan hastalarda afektif mizaçları araştıran tek bir çalışma yapılmıştır. Bu çalışmada afektif mizaç ve postpartum depresif bozukluğun varlığı eş zamanlı olarak değerlendirilmiş ve siklotimik ve endişeli mizacın psikososyal faktörlerden bağımsız olarak önemli risk faktörleri arasında olduğu bildirilmiştir¹⁷. Çalışmamızın metodolojisi Masmoudi ve ark.’nın²⁶ araştırmasından farklıydı. İlk olarak, gebeliğin üçüncü trimesterinde, mevcut veya daha önce psikiyatrik bozukluğu olmayan kadınlarda afektif mizaçları değerlendirdik. İkinci olarak çalışmamız, hem gebelik hem de doğum sonrası dönemde hastaları inceleyen bir takip çalışmasıdır. Sonuçlarımız ilk olarak, doğum sonrası depresif bozukluğa eğilim geliştirmeye aday olarak siklotimik ve endişeli mizaçları göstermektedir. Ayrıca çocuk sahibi olmak ve doğum yeri postpartum depresif bozukluk için diğer risk faktörleri olarak bulunmuştur.

Çalışmanın Kısıtlılıkları

Bu çalışma için güç analizi yapmamıza rağmen, postpartum depresif bozukluğun yordayıcılarının belirlenmesi açısından genel bir sonuca varmak için hasta sayılarının küçük olduğu düşünülebilir. Bu, çalışmamızın en büyük sınırlamasıdır. Psikiyatrik bozukluğu olmayan gebeler de dahil olmak üzere, aynı trimesterdeki hastaların değerlendirilmesi, PPD’nin ilk dört hafta içinde değerlendirilmesi ve çalışmamızın prospektif dizaynı çalışmamızın güçlü yönleri arasında sayılabilir.

CONCLUSION

In conclusion, we argue that cyclothymic and AT can be considered candidates for predicting tendency to PPD beside other risk factors. Pregnant women who have higher scores specifically for cyclothymic and AT should be evaluated in terms of affective disorders.

Ethics

Ethics Committee Approval: It was approved by Tekirdağ Namık Kemal University Non-Invasive Clinic Research Ethical Committee (date and approval number: 2018/124/08/15).

Informed Consent: Consent form was filled out by all participants.

Peer-review: Externally peer-reviewed.

Authorship Contributions

Surgical and Medical Practices: E.A., E.B., Concept: E.A., E.B., Design: E.A., E.B., Data Collection or Processing: E.A., E.B., Analysis or Interpretation: E.A., E.B., Literature Search: E.A., E.B., Writing: E.A., E.B.

Conflict of Interest: No conflict of interest was declared by the authors.

Financial Disclosure: The authors declared that this study received no financial support.

References

Kessler RC. Epidemiology of women and depression. J Affect Disord. 2003;74:5-13.

American Psychiatric Association: Diagnostic and Statistical Manual of Mental Health Disorders: DSM-5 Vol. American Psychiatric Publishing. 2013, Washington, DC.

Brummelte S, Galea LA. Postpartum depression: Etiology, treatment and consequences for maternal care. Horm Behav. 2016;77:153-66.

Patel M, Bailey RK, Jabeen S, Ali S, Barker NC, Osiezagha K. Postpartum depression: a review. J Health Care Poor Underserved. 2012;23:534-42.

Cohen LS, Wang B, Nonacs R, Viguera AC, Lemon EL, Freeman MP. Treatment of mood disorders during pregnancy and postpartum. Psychiatr Clin North Am. 2010;33:273-93.

Norhayati MN, Hazlina NH, Asrenee AR, Emilin WM. Magnitude and risk factors for postpartum symptoms: a literature review. J Affect Disord. 2015;175:34-52.

Dubber S, Reck C, Müller M, Gawlik S. Postpartum bonding: the role of perinatal depression, anxiety and maternal–fetal bonding during pregnancy. Arch Womens Ment Health. 2014;18:187-95.

Darcy JM, Grzywacz JG, Stephens RL, Leng I, Clinch CR, Arcury TA. Maternal depressive symptomatology: 16-month follow-up of infant and maternal health-related quality of life. J Am Board Fam Med. 2011;24:249-57.

Beck CT. Predictors of postpartumdepression: An Update. Nurs Res. 2001;50:275-85.

Rubertsson C, Wickberg B, Gustavsson P, Rådestad I. Depressive symptoms in early pregnancy, two months and one year postpartum-prevalence and psychosocial risk factors in a national Swedish sample. Arch Womens Ment Health. 2005;8:97-104.

Banti S, Mauri M, Oppo A, Borri C, Rambelli C, Ramacciotti D, et al. From the third month of pregnancy to 1 year postpartum. Prevalence, incidence, recurrence, and new onset of depression. Results from the perinatal depression-research & screening unit study. Compr Psychiatry. 2011;52:343-51.

Rasmussen MH, Strøm M, Wohlfahrt J, Videbech P, Melbye M. Risk, treatment duration, and recurrence risk of postpartum affective disorder in women with no prior psychiatric history: A population-based cohort study. PLoS Med. 2017;14:e1002392.

Rihmer Z, Akiskal KK, Rihmer A, Akiskal HS. Current research on affective temperaments. Curr Opin Psychiatry. 2010;23:12-8.

Vázquez GH, Tondo L, Mazzarini L, Gonda X. Affective temperaments in general population: a review and combined analysis from national studies. J Affect Disord. 2012;139:18-22.

Yazici E, Terzi H, Bosgelmez S, Yazici AB, Zincir SB, Kale A. Affective temperaments in pregnancy. Gynecol Endocrinol. 2014;30:894-8.

Yazici E, Uslu Yuvaci H, Yazici AB, Cevrioglu AS, Erol A. Affective temperaments during pregnancy and postpartum period: a click to hyperthymic temperament. Gynecol Endocrinol. 2018;34:265-9.

Vahip S, Kesebir S, Alkan M, Yazici O, Akiskal KK, Akiskal HS. Affective temperaments in clinically-well subjects in Turkey: initial psychometric data on the TEMPS-A. J Affect Disord. 2005;85:113-25.

Engindeniz AN, Küey L, Kültür S. Edinburgh doğum sonrası depresyon ölçeği Türkçe formu geçerlilik ve güvenilirlik çalışması. Bahar Sempozyumları 1. Kitabı, Psikiyatri Derneği Yayınları, Ankara 1997. p. 51-2.

Giallo R, Cooklin A, Wade C, D’Esposito F, Nicholson JM. Maternal postnatal mental health and later emotional-behavioural development of children: the mediating role of parenting behaviour. Child Care Health Dev. 2014;40:327-36.

Goodman SH, Rouse MH, Connell AM, Broth MR, Hall CM, Heyward D. Maternal depression and child psychopathology: a meta-analytic review. Clin Child Fam Psychol Rev. 2011;14:1-27.

McMahon CA, Barnett B, Kowalenko NM, Tennant CC. Maternal attachment state of mind moderates the impact of postnatal depression on infant attachment. J Child Psychol Psychiatry. 2006;47:660-9.

Akiskal HS. [The depressive patient before depression]. Encephale. 1992;4:485-9.

Kelsoe JR. Arguments for the genetic basis of the bipolar spectrum. J Affect Disord. 2003;73:183-97.

Akiskal KK, Akiskal HS. The theoretical underpinnings of affective temperaments: implications for evolutionary foundations of bipolar disorder and human nature. J Affect Disord. 2005;85:231-9.

Röttig D, Röttig S, Brieger P, Marneros A. Temperament and personality in bipolar I patients with and without mixed episodes. J Affect Disord. 2007;104:97-102.

Masmoudi J, Trabelsi S, Charfeddine F, Ben Ayed B, Guermazi M, Jaoua A, et al. Evaluation des tempéraments affectifs dans la symptomatologie dépressive du post-partum [Evaluation of affective temperaments in the postpartum depressive symptomatology]. Encephale. 2010;36:14-21.