You are on page 1of 6

Obstetric, Somatic, and Demographic Risk Factors for Postpartum Depressive Symptoms

Ann Josefsson, MD, Lisbeth Angelsio ¨o ¨ , MD, Go ¨ ran Berg, MD, PhD, Carl-Magnus Ekstro ¨ m, MD, ¨ , PhD Christina Gunnervik, MD, Conny Nordin, MD, PhD, and Gunilla Sydsjo
OBJECTIVE: To identify and test the predictive power of potential independent risk factors of postpartum depressive symptoms during pregnancy and the perinatal period. METHODS: We conducted a case-control study where 132 women with postpartum depressive symptoms were selected as an index group and 264 women without depressive symptoms as a control group. Data related to sociodemographic status, medical, gynecologic, and obstetric history, pregnancy, and perinatal events were collected from standardized medical records. RESULTS: The strongest risk factors for postpartum depressive symptoms were sick leave during pregnancy and a high number of visits to the antenatal care clinic. Complications during pregnancy, such as hyperemesis, premature contractions, and psychiatric disorder were more common in the postpartum depressed group of women. No association was found between parity, sociodemographic data, or mode of delivery and postpartum depressive symptoms. CONCLUSION: Women at risk for postpartum depression can be identified during pregnancy. The strongest risk factors, sick leave during pregnancy and many visits to the antenatal care clinic, are not etiologic and might be of either behavioral or biologic origin. The possibilities of genetic vulnerability and hormonal changes warrant further investigation to reach a more thorough understanding. (Obstet Gynecol 2002;99:223– 8. © 2002 by the American College of Obstetricians and Gynecologists.)

challenge a woman’s mental health. Postpartum depression, which often resembles other forms of major depression, affects 10 –20% of all mothers.2–11 It may have a deleterious effect on the woman’s social and personal adjustment, the marital relationship, and the motherinfant interaction. Furthermore, there is a 30 –50% risk of relapse of depression in a future pregnancy.1,12 Maternal depression early in the infant’s life may affect the child’s psychologic development with significant intellectual deficits as a result.13–15 Other consequences for the child include higher risk of accidents, sudden infant death syndrome, and a higher frequency of hospital admissions.16 –18 Various explanatory models on the etiology have been proposed; probably postpartum depression is a result of an interaction between genetic vulnerability, hormonal changes, and major life events.19 –22 Recent studies have focused on psychosocial stressors and previous psychiatric history in a woman’s life as major risk factors for developing postpartum depression.23–27 The literature concerning obstetric and perinatal risk factors is sparse and shows little concordance.2,8,28,29 The hypothesis of this study was that complications during pregnancy, delivery, and/or the perinatal period are associated with an increased risk of postpartum depression. MATERIALS AND METHODS The Swedish antenatal health care system reaches almost 100% of all pregnant women (Swedish National Board of Health and Welfare). The antenatal care clinics provide regular check-ups on the physical and psychologic health during pregnancy and puerperium. The same percentage, as above, is valid for deliveries as there are no private maternity hospitals in Sweden and home deliveries are rare. The original sample in the present study comprises the total population of pregnant women consecutively registered at the antenatal care clinics in four communities in the southeast region of Sweden.11 Enrollment took place in separate 3-month periods for each of the four commu-

Epidemiologic data from around the world show that depression is approximately twice as common in women than men and that its first onset peaks during the childbearing years.1 Pregnancy, miscarriage or fetal death, infertility, and the postpartum period may especially
From the Department of Health and Environment, Division of Obstetrics and Gynaecology, and Department of Neuroscience and Locomotion, Division of Psychiatry, Faculty of Health Sciences, University of Linko ¨ ping, Linko ¨ ping, Sweden; Department of Obstetrics and Gynaecology, Va ¨ rnamo, Sweden; Department of Obstetrics and Gynaecology, Norrko ¨ ping, Sweden; and Department of Obstetrics and Gynaecology, Kalmar, Sweden. This study was supported by grants from The Health Research Council in the southeast of Sweden.

VOL. 99, NO. 2, FEBRUARY 2002 © 2002 by The American College of Obstetricians and Gynecologists. Published by Elsevier Science Inc.

0029-7844/02/$22.00 PII S0029-7844(01)01722-7


No differences were found between women with or without a depressive symptomatology. and obstetric history. were significantly more prone to develop depressive symptoms postpartum. actual chronic medical diseases. The Edinburgh Postnatal Depression Scale is a 10-item self-report scale. the cutoff level of 10 was used as the dependent variable. pregnancy. As control group. The study was approved by the Regional Ethics Committee for Human Research of the Faculty of Health Sciences. Women with an earlier history of two or more abortions and a history of earlier obstetric complications. delivery. and the findings were identical or similar to earlier studies. One woman in the control group delivered a dead infant and was therefore excluded. such as hyperemesis. The eligible women were approached in gestational weeks 35–36 and received both written and oral information from their midwife before giving consent. two with anxiety. medical. suicidal ideas. were calculated for categoric variables. The following data were collected: age. and neonatal records. no differences were found. The data were manually extracted from the records by the main author (AJ). and is well accepted by the women and the staff. The prevalence of depressive symptoms. medical records from earlier deliveries were also scrutinized. Statistical significance was defined as two-sided P values using a significance level of 5%. Five of the items are concerned with dysphoric mood. and perinatal events were obtained. In multiparas.nities during 1997–1999. all women with depressive symptoms on the Edinburgh Postnatal Depression Scale at 6 – 8 weeks and/or 6 months postpartum from two of the four communities were selected as an index group (n ϭ 132). psychiatric disorder. Linko ¨ ping University (No.30 Each item is scored on a 4-point scale (0 –3). All analyses were done using the SPSS program 10. RESULTS Distributions of independent variables are shown in Tables 1–5. Records from three index women and two control women were not found. Concerning delivery data.32 Validity of the Swedish version has been tested. Odds ratios. presented with 95% confidence intervals. and one each with guilt.. Women with postpartum depressive symptoms stayed longer at the hospital after delivery but did not have a higher frequency of complicated puerperium than nondepressed women (Table 5). was 13% at the 6 – 8 weeks assessment and unaltered at 6 months postpartum. and the puerperium were registered in the standardized and identical Swedish antenatal. the sensitivity for the detection of major depression was almost 100% and the specificity 82%. As the women included in this study were enrolled in pregnancy weeks 35–36. IL). In Table 2. parity.9.” The Edinburgh Postnatal Depression Scale has been translated into at least 11 languages. Number of visits at the antenatal care clinics and number of women on sick leave were also more frequent in this group (Table 3). and are thus prospectively related to the development of depressive symptoms. and number of induced abortions. takes only a few minutes to complete. 264 women without depressive symptomatology on the Edinburgh Postnatal Depression Scale were randomly chosen from all four communities. delivery. number of visits at the antenatal care clinics before delivery (midwife and physician). respectively. sick leave during pregnancy. the minimum and maximum total score ranging from 0 to 30. and neonatal data. or obstetric complications. The explanatory variables were sociodemographic data. Women who gave birth to babies with congenital major malformations were also more de- 224 Josefsson et al Postpartum Depressive Symptoms OBSTETRICS & GYNECOLOGY . we describe earlier medical history with focus on gynecologic and obstetric data. Sixty-nine (4. The Edinburgh Postnatal Depression Scale cannot confirm a diagnosis of depressive illness. and psychiatric disorder.10 Cox et al30 proposed a cutoff level of 10 if the test is to be used for screening purposes in the postpartum period. gynecologic. there are no premature deliveries in the sample. The sociodemographic variables are shown in Table 1. but when selecting this threshold. delivery. occupation. In this study.1 (SPSS Inc. The dependent variables were depressed or nondepressed. pregnancy complications. or extrauter- ine pregnancies. an Edinburgh Postnatal Depression Scale score of 10 or more. Any history of infertility. 97133).4%) women declined to participate in the original study. A total of 1558 women were approached. Complications during the present pregnancy (Table 3). marital status. except from intrapartal analgesia with morphine analogues.31 including Swedish. Chicago. and “not coping. presented in Table 4. Out of 1489 women. specifically designed to screen for postpartum depression in community samples. mainly acute cesarean section and instrumental delivery. miscarriages. The prevalence of depressive symptoms did not differ between the communities within each assessment.33 The Edinburgh Postnatal Depression Scale is easy to administer. Differences were tested with Student t test for normally distributed continuous variables and the MannWhitney U test for continuous variables not normally distributed. were more common in the postpartum depressed group of women. The scale rates the intensity of depressive symptoms present within the previous 7 days. Logistic regression with conditional stepwise backward elimination was used when multiple variables were considered simultaneously.11 All data related to the pregnancy. premature contractions.

73 (0.23.8 0. 1.94 (0.01) 86.3 4.77.0 6.50) 1. 2. * Student t test. 2. Values are given as %.03) 1.29 (0.6 0. * A woman can have more than one chronic disease.71) 2.12 (0. 6.5 0. 2.0 Nondepressed 12.766.2 4.9 4.6 1. 19.3 98.3 17.90) 2.3 Unadjusted OR (95% CI) 1. 1.1 3.0 10. 7.2 5.9 7.0 1.1 97. 4. 11.30 (0. Values are given as %.51) 1 (reference) 2.5 14. 2.6 35.67) 1.8 5.28.8 10. 14.40.2 16.04 (0. 1.67) 1 (reference) 2. 2.56) 1 (reference) 1.20) 5. 1.8 7.29) 1.02 (0.35) 1.7 25. 1.30 (0.1 7.68.17 (0. 2.9 4. † A woman can have more than one obstetric complication.70.2 Unadjusted OR (95% CI) 1.3 5.66 (0.3 89. Subsequent multivariable analysis is shown in Table 6.2 45.58. unless otherwise stated. Only women with earlier deliveries are analyzed. 3.10 (0.57) Abbreviations as in Table 1.20.37. 5.85 (1. multivariable analyses were per- Depressed 13. Sick leave during pregnancy and number of visits at the antenatal Table 2. 2. Sociodemographic Data Variable Age (y) 16–24 25–34 35–46 Mean age (y)* Parity 0 1–2 Ն3 Marital status Solitary Married/cohabiting Socioeconomic groups Unskilled workers Skilled workers Lower white collar workers Middle/high white collar workers and self-employed Students Unemployed Others OR ϭ odds ratio.54 (0.83 (0.6 24. unless otherwise stated.50 (0.4 7.7 86.25.88 (0. 29.88) 1. 4.8 13.56) 1 (reference) 0.2 2.90) 1.7 1.04 (0.95) 1 (reference) 1.3 29.6 71.79 (1. 2.55.5 41.79) 4.39.1 13.66.7 20.74.5 25.7 6.6 15. 4.13) 1 (reference) 0.7 45. Depressed 15. Gynecologic/Obstetric History Variable Chronic medical disease including drug addiction* Psychiatric disorder Induced abortions 0 1 Ն2 Miscarriages 0 1 Ն2 Extrauterine pregnancies 0 1 Ն2 Infertility (Ͼ1 y) Obstetric complications† No complications Acute cesarean section Elective cesarean section Instrumental delivery Perineal tears/excessive bleeding intrapartum Premature delivery Preeclampsia care clinic are the variables with the strongest association to postpartum depressive symptoms for all women.1 80. 99.7 29.32 (0.19 (0.7 8.63.18 (0.7 6.54) 1.19 (0. P ϭ . 1.07 (0.91.61. VOL.9 38.Table 1.0 8.5 68.4 17. CI ϭ confidence interval.85 (1.6 7.69.9 Nondepressed 9.45. NO.3 9.65) 1.11) 1 (reference) 0.3 97.47. To facilitate comparisons. 6.3 64.6 18.09 (0.09.0 15.66 (0.3 13.4 2.37) 1 (reference) 1.8 2.69.8 10.9 14.8 19.60.29) 0.1 7.3 97.28) 1. FEBRUARY 2002 Josefsson et al Postpartum Depressive Symptoms 225 .25) pressed than the others were (Table 5). 5.17.

Delivery-Related Data Variable Induction of labor Normal delivery Instrumental delivery Acute cesarean section Elective cesarean section Perineal tears and excessive bleeding intrapartum Number of hours from established contractions to delivery* Number of hours of pushing† Paracervical blockade Pudendal blockade Morphine analogues N2 O Epidural anesthesia Abbreviations as in Table 1.10) 2. † Student t test.37 (0.9 5.59) 2.21.72 (0.53. 2.1 17.1 10.35. * Student t test.37.8 22.55.2 31.7 4. unless otherwise stated.0 5. Other abbreviations as in Table 1.61 ( 19.01 (0. A woman can have more than one delivery-related variable.03) 1. 13.5 68.97 (1.866.3 57.09.87) 226 Josefsson et al Postpartum Depressive Symptoms OBSTETRICS & GYNECOLOGY .551.7 4.7 8. 5.10 (0.6 2.61) 1. Values are given as %.9 62.94 (1.8 17.29) ANC ϭ antenatal care.1 4. Values are given as %. 4.0 80. 7.4 0.3 37. 14. 2. 6.3 58. 1.63 (2. 2.04 (0.41) 1. § Preeclampsia.76) 2. Depressed 10.9 9.4 0. 5.29) 1. and miscellaneous. Pregnancy-Related Data Variable Visits to midwife at the ANC Յ8 visits 9–12 visits Ն13 visits Visits to physician at the ANC Յ2 visits Ն3 visits Smoking No Yes Admission to hospital because of pregnancy complications Number of days* Pregnancy complication† No complications Preeclampsia Hyperemesis Back pain Vaginal bleeding Premature contractions Miscellaneous Psychiatric disorder Delivery in gestational week‡ Reason for sick leave during pregnancy No sick leave Pregnancy-related complications§ Psychiatric disorder࿣ Miscellaneous Depressed 33.9 36. 3. 1.36) 1. 2. 2.81 (0.470.1 8.9 1. unless otherwise stated.19) 1 (reference) 0. 2. back pain.2 Nondepressed 9.33. Table 4.04) 0.1 6. 2.31) 2. 5.7 57.9 5.24 (0. 2.20 (0.2 10. 3. 2.93 (1.4 64.0 Unadjusted OR (95% CI) 17.2 40.2 7.58 0.0 12.4 Unadjusted OR (95% CI) 1 (reference) 1.85. 14.67 (1.1 47.2 1.7 76.13) 1.70 (0. P ϭ . P ϭ .66 (0.60 (0.18) 1 (reference) 0.1 8. 2.3 6.2 2. † A woman can have more than one pregnancy complication.21 (0.3 1.8 93. * Mann-Whitney U test.01.21.9 40.31 ( (1.82) 1 (reference) 2.58 (0. 145.41) 1.66 (0.9 15.8 (1.6 57.72 (1.81) 1 (reference) 1. P ϭ .2 90.48) 5.31) 1.1 7.2 69.3 4.51 4.4 Nondepressed 47.0 10. hyperemesis.68) 1.90.3 1.23 0.0 10.9 5. 2.3 12.49 (1.52 3. American Psychiatric Association.48) 2.60.8 6.9 33. vaginal bleeding.2 1. 56.65.8 10.32 (0. P ϭ .45.Table 3.69) 2.88. ‡ Student t test.0 6. ࿣ Psychiatric disorders listed in DSM-IV.99.2 10. premature contractions.4 7.6 36.35) 5.

9 3589 21. respectively.09 (1. Depressed 9. The strongest risk factors.34 but not with some earlier studies. 1.0 Nondepressed 11. Bond A. Watson JP. We found no association between delivery complications and the development of depressive symptoms. The possibilities of genetic vulnerability and hormonal changes ought to be investigated further to reach a more thorough understanding. A prospective study of emotional disorders in childbearing women. Most of the predictors found were obvious already during pregnancy. REFERENCES 1. 2. a potential weakness in this study might be an underestimation of an earlier or ongoing psychiatric condition influencing the results.60) 5.17.35 Our results show that women at risk for postpartum depression can be identified during pregnancy. Kumar R.5 0.4 10.269: 799 – 801. † 0 ϭ Յ 10 visits. Br J Psych 1984. VOL.50) 1 (reference) 1. * Student t test.3 3. 2. 27. The absence of correlation between sociodemographic variables and depressive symptoms is in line with previous studies. Robson K.868.98 (1.19) P . However.0 16. Day A. NO.7 62.7 97. Elliott SA.03. this result supports our hypothesis that pregnancy complications are risk factors for postpartum depression. which have reported a strong link between cesarean section and postpartum depression. 1 ϭ Ն 11 visits (midwife ϩ physician).3 91. Olfson M. gynecologic. An advantage of this study is the fact that we had reached the total pregnant population at the time of the study and that all data were extracted from standardized medical records and not from maternal recall. unless otherwise stated.70) 0.29. 4. 3. 4. Rugg AJ.1 93. Hyperemesis and premature contractions were more common in the postpartum depressed group of women. 2. FEBRUARY 2002 Josefsson et al Postpartum Depressive Symptoms 227 .9 59.144: 35– 47. DISCUSSION In this population-based case-control study.86.01. 2.6 Unadjusted OR (95% CI) 0. P Ͻ . Weissman MM. To our knowledge. Nevertheless.001.82) formed for multiparas and primiparas. * Multiple logistic regression. Campbell EA. Non-psychotic psychiatric disorder after childbirth: A pro- Abbreviations as in Tables 1 and 3. 1. sick leave and a high number of visits at the antenatal care clinics. Kennerly H. and obstetric history.23. which could facilitate planning of prevention and intervention in the future.39) 2.79 (0.28. Science 1995.008 . Multivariable Analysis*: All Women (n ϭ 390) Variable Number of visits at the ANC† Sick leave during pregnancy OR (95% CI) 1.8 3581 29. The strongest risk factors for postpartum depressive symptoms were sick leave during pregnancy and a high number of visits to the antenatal care clinic.144:453– 62. but no further information was gained. are not etiologic and might be of either behavioral or biologic origin.92) 1.31 but whether this is on a psychosocial basis or a result of hormonal and genetic vulnerability remains to be investigated.36.46. 3.39. Depression in women: Implications for health care research.44 (0.3 96. P ϭ . Cooper PJ. information regarding a wide range of potential risk factors was collected. Brough DI. This is consistent with the findings in two recent studies. Psychiatric disorder in pregnancy and the first postpartum year. 4. this is the only study that includes earlier medical. These risk factors did not remain significant on the multivariable level.26 (1.3 The risk factors identified in the gynecologic and obstetric history were two induced abortions or more and among multiparas a history of acute cesarean section or instrumental delivery in the past. Antenatal depressive symptoms and postpartum depression are correlated.Table 5. Perinatal Data Variable Neonatal morbidity Congenital major malformation Birth weight* (g) Postnatal hospital stay Յ2 days 3–5 days Ն6 days Breast-feeding Puerperium without complications Abbreviations as in Table 1.005 pregnant woman.85 (1. Br J Psych 1984.26) 1.2.46 (0.82 (0. 99. Values are given as %. The reasons for sick leave were mainly psychiatric disorders and pregnancy-related complications.11. Explanatory models for these symptoms might be somatization of pregnancy-related anxiety or depression but might also be an effect of hormonal changes in the Table 6.

Br J Psych 1987. Zekoski E. The impact of postpartum depression on boys’ intellectual development. Nielsen Forman D. Linko ¨ ping. Sweden. psychosocial and medical risk factors. Bousquet A. E-mail: ann.4: 29 – 47. 14. Rates and risk of postpartum depression — A meta-analysis. Ziv A. Hedegaard M. Saisto T. Murray L. Allen H. Sandler M. course and nature. 2001. Bydlowski M. Postpartum depession in an Israeli cohort: Demographic. Papiernik E. 29. Course of mood and anxiety disorders during pregnancy and the postpartum period. psychiatric disorder of mother. Depression during pregnancy and the puerperium. 2001. Motherhood and mental health. Conne-Perre ´ ard E.8: 37–54. Berg G. 80:39 – 45. Kumar R. Br J Obstet Gynaecol 2000. Address reprint requests to: Ann Josefsson. Appleby L. 1994. Righetti-Veltema M. 28(Suppl 1):S13–S16. Flanagan C. 98:272–5. Lund W. Br J Psych 1996. 26. Accepted October 18. Links between early post-partum mood and post-natal depression. Mothers’ depression and their children’s attendance at medical facilities. Thomas R. Risk factors and predictive signs of postpartum depression. Monsen Bo ¨ rjesson K. Pawlby Nemeroff CB. Br J Psych 1989. Br J Psych 1994. Glover V. Lusky A. Faragher B.152:799 – 806.80:251–5. 30. and accidents to children.292:1165–7. Warner R. 11. Harris B. 20. Hannah P. Increased risk of postpartum depression after emergency caesarean section. Todd AL. Videbech P. Brown GW. Fossey L. Bågedahl-Strindlund M. Wickberg B. Schmucker G. Postpartum depression: Identification of women at risk. Stowe ZN. MD. J Affect Dis 1998. 18. Received June 5. In: Kumar R. Cohen LS. 7. Med J Aust 1992. somatic growth and social outcome. Holden J. Adams D. Bågedahl-Strindlund M. Whitton A. Course and correlates of postpartum depression during the transition to parenthood. Impact of maternal postpartum depression on cognitive development of young children. J Paediatr Child Health 1992. Sharp D. Received in revised form September 17. Acta Psych Scand 1996. Cogill SR. Fung H. Motherhood and mental illness 2. Barell V. Prevalence of depressive symptoms in late pregnancy and postpartum. SE-581-85. Postnatal depression and SIDS: A prospective study.166:191–5.21:99 –108. 25. Popper S. Huckle P. BMJ 1986.29: 579 – 82. Course and recurrence of postpartum depression: Evidence for the specificity of the diagnostic concept. Zekoski EM. Sagovsky R. Caplan HL.59(Suppl 2):29 –33. Scand J Soc Med 1988. O’Hara MW. Wolkind S.107:1210 –7. Psychosocial predictors of disappointment with delivery and puerperal depression. Postpartum depression: A hidden illness. 28. Philipps LA. Dale FH. Controlled prospective study of postpartum mood disorders: Comparison of childbearing and non-childbearing women. Lancet 1978. 12.36: 1315–36. Campbell SB. Biological and hormonal aspects of postpartum depressed mood. Gyllang C. University of Linko ¨ ping. 15. London: Gaskell. Postpartum depression. Mitchell EA. 13. Sydsjo ¨ G. 10. Hwang CP. Swain AM. 27. Harris B. Thompson JM. J Clin Psych 1998.164:288 –92. J Psychosom Obstet Gynecol 2000. Social class. Holden JM.49:167– 80. Br J Psych 1992. Cooper PJ.150:782– 6. 9. 33. Stewart AW.5. eds. J Child Psychol Psych 1995. et al. Wright EJ. Taylor BJ. Manzano J. Scand J Caring Sci 1993. Detection of postpartum depression: Development of the 10-item Edinburgh Postpartum Depression Scale. Glasser S.16:121–7. Nurmi JE. Dalby Salvig J. 2001. Nordin C. Br J Psych 1995. 24. Go ¨ teborg University. Robson K. J Abnorm Psychol 1990. Altshuler LL. 1988. spective study of prevalence. Johns S. Br J Psych 1988. Shoham A. Postpartum blues: A clinical syndrome and predictor of postpartum depression? J Psychosom Obstet Gynecol 1997. Boyko V. Josefsson A. Use of the Edinburgh Postpartum Depression Scale in some Swedish child health care centers. Secher NJ. The Edinburgh Postpartum Depression Scale: Validation on a Swedish community sample. The use of rating scales to identify postpartum depression. Cox J. 23. 19. Dev Psychopathol 1992. Int Rev Psych 1996. 16. Thesis. Salmela-Aro K.94:181– 4. 228 Josefsson et al Postpartum Depressive Symptoms OBSTETRICS & GYNECOLOGY . Alexandra H. Meyera T. Halmesma ¨ ki E. Perinatal psychiatry: Use and misuse of the Edinburgh Postpartum Depression Scale. University Hospital. 6. Children of mentally ill mothers: Mental development. J Clin Psych 1997. Postpartum depression: A comprehensive review. et al. josefsson@lio. 58(Suppl 15):26 –32.18:17–21. Division of Obstetrics and Gynaecology. London: Wright. Cohn JF. Llewellyn AM. O’Hara MW. Brockington JF. incidence. Lee A. eds. Demographic and obstetric risk factors for postpartum psychiatric morbidity. 157:172– 4. 35. 31. 22. 1996. Hassall IB. 34.168:607–11. Kumar R. 8. Wickberg B. 17.154:813–7. Hendrick V.99:3–15. Acta Obstet Gynecol Scand 2001. O’Hara MW.160:777– 80. Webster ML. Acta Psych Scand 1998. Boyce PM. Brockington I.1: 378 – 81. Oxford: Oxford University Press. 1996. Davidson S. J Psychosom Res 1985. 21.7:149 –54. Cox JL. 32. Acta Obstet Gynecol Scand 2001.