Older first-time mothers and early postpartum depression: a prospective cohort study of women conceiving spontaneously or with assisted reproductive technologies

Page 1

Received

2011; pu

C.A.M. ha

nothing

to disclo

IVF Aust

Funded th

financia

Melbou

Reprint re

Health D

NSW, 21

FC

1218

Older first-time mothers and early postpartumdepression: a prospective cohort study of womenconceiving spontaneously or with assistedreproductive technologies

Catherine A. McMahon, Ph.D.,a Jacky Boivin, Ph.D.,b Frances L. Gibson, Ph.D.,c Jane R. W. Fisher, Ph.D.,d

Karin Hammarberg, Ph.D.,e Karen Wynter, Ph.D.,d and Douglas M. Saunders, M.D.f,g

a Centre for Emotional Health, Department of Psychology, Macquarie University, North Ryde, New South Wales, Australia;b School of Psychology, Cardiff University, Cardiff, Wales, United Kingdom; c Institute of Early Childhood Macquarie

University, North Ryde, New South Wales, Australia; d Jean Hailes Research Unit, School of Public Health and Preventive

Medicine, Monash University, Clayton, Victoria, Australia; e Centre for Women’s Health, Gender and Society, Melbourne

School of Population Health, University of Melbourne, Melbourne, Victoria, Australia; f IVF Australia, St. Leonard’s; andg University of Sydney, Sydney, New South Wales, Australia

Objective: To evaluate whether older first-time mothers (R37 years) have higher rates of postpartum depressioncompared with younger first-time mothers, controlling for mode of conception and known risk factors forpostpartum depression.Design: Prospective cohort study.Setting: Assisted reproductive technology (ART) clinics in two large Australian cities and public and privateantenatal clinics and/or classes in the vicinity of ART clinics.Patient(s): Nulliparous women who had conceived spontaneously (n ¼ 295) or through ART (n ¼ 297) in threeage-groups: younger, 20 to 30 years (n ¼ 173); middle, 31 to 36 years (n ¼ 214); and older,R37 years (n ¼ 189).Intervention(s): Semistructured interviews and questionnaires.Main Outcome Measure(s): Major depressive disorder in the first 4 months after birth as assessed by structureddiagnostic interview.Result(s): The study performed 592 complete pregnancy assessments and 541 postpartum assessments. Theprevalence of major depressive disorder was 7.9%, at the lower end of community rates. Neither maternalage-group nor mode of conception was statistically significantly related to depression.Conclusion(s): Older first-timemothers, whether conceiving through ARTor spontaneously, do not show increasedvulnerability to postnatal depression. (Fertil Steril� 2011;96:1218–24. �2011 by American Society forReproductive Medicine.)

Key Words: Assisted reproductive technology, maternal age, postpartum depression, psychosocial

The trend for women to have their first baby in their late 30s or 40s is have more problematic adjustment, including a greater risk of post-

now well established in developed countries. Reasons includea range of interacting social, educational, employment, and eco-nomic factors, many of which seem unlikely to change. Indeed,some argue that delayed first birth is likely to become the normand that health practitioners will increasingly see older first-timeparents (1). The biological complications of delayed childbearing(fertility related and obstetric) are well documented (2), but less isknown about the psychosocial implications, particularly for adjust-ment during the early postpartum period. Anecdotal evidence sug-gests that older mothers are perceived by health professionals to

July 11, 2011; revised August 22, 2011; accepted August 25,

blished online October 1, 2011.

s nothing to disclose. J.B. has nothing to disclose. F.L.G. has

to disclose. J.R.W.F. has nothing to disclose. K.H. has nothing

se. K.W. has nothing to disclose. D.M.S. is a former Director at

ralia, an IVF clinic that contributed to funding.

rough an Australian Research Council (ARC) Linkage Grant with

l contributions from private IVF clinics (IVF Australia and

rne IVF) as linkage industry partners.

quests: Catherine A. McMahon, Ph.D., Centre for Emotional

epartment of Psychology, Macquarie University, North Ryde,

09, Australia (E-mail: [email protected]).

ertility and Sterility� Vol. 96, No. 5, November 2011opyright ª2011 American Society for Reproductive Medicine, P

partum depression, but empirical evidence has been limited (3).Conception, pregnancy, and childbirth are more medically com-

plicated for older women due to their greater likelihood of requiringassisted reproductive technology (ART) (1) and donated eggs orembryos, and their increased incidence of multiple pregnancies,gestational complications, and cesarean birth (4). This high-riskcontext, particularly after ART conception, may be associatedwith elevated anxiety about the pregnancy outcome (5); maternalmood during pregnancy is now recognized as one of the strongestpredictors of postpartum depression (6). Cesarean birth is associatedwith maternal physical morbidity, a negative impact on the firstpostnatal contact with the infant, and possible adverse effects onbreastfeeding capacity and maternal mood (7).

Potentially offsetting these biological risk factors are the morefavorable socioeconomic characteristics of older mothers (8) (e.g.,higher education levels, more financial security) together with fac-tors relating to maturity (9) (e.g., wisdom, stability of relationship)that are protective against postpartum depression in the short term(6) and are widely recognized as protective for parenting and childdevelopment in the longer term (10). Older mothers, however,may receive less peer support, as many of their contemporarieshave completed their childbearing (11).

0015-0282/$36.00ublished by Elsevier Inc. doi:10.1016/j.fertnstert.2011.08.037

Page 2

The interactive effects of these risk and protective factors arelikely to be complex, and there may be a point where biologicalrisk factors limit any psychosocial benefits of childbearing at anolder age (12). The limited evidence to date regarding relations be-tween maternal age and depression in the context of parenting hasbeen equivocal, with a negative association during pregnancy(13, 14), a positive association at child ages 4 to 10 years (15),and a curvilinear association with depression rated retrospectively(12). Assessment at different times in the parenting cycle may con-tribute to these inconsistent findings, and none to date have focusedon the early postpartum period.

Further, while age-related declines in fertility are well recog-nized, studies to date of maternal age and depression have failedto take account of the potentially interactive effects of age, infertil-ity, and ART treatment history. Some studies of depression have ex-amined age effects but not mode of conception (9). Some haveconsidered age differences but only in women conceiving throughART (13, 15), whereas others have examined depression inwomen conceiving through ART (16) but have not considered age.A recent systematic review concluded there was no evidence for in-creased perinatal depression rates in mothers conceiving throughART, but it also noted significant methodologic limitations in thestudies to date, including inadequate sample sizes, lack of appropri-ate comparison groups, and failure to control for known socioeco-nomic risk and protective factors for depression (17). Inconsistentfindings may also be due to different measures (symptom check lists,diagnostic interviews, behavioral indexes of stress/coping). Studiesinvestigating service-use as an index of parental postpartum distresshave, for example, demonstrated that women conceiving throughARTwere three to four times more likely than those who conceivedspontaneously to be admitted to intensive parent support services inAustralia, but the investigators acknowledged that this could beattributable, at least in part, to older maternal age, cesarean birth,or multiple birth (16, 18).

The aim of the multicenter Parental Age and Transition to Parent-hood Australia (PATPA) study was to investigate how age at firstbirth and mode of conception were associated with adjustment toearly parenthood. Notwithstanding equivocal results to date, wetested the proposition that older first-time mothers have higher ratesof postpartum depression. We also proposed that the relationshipbetween maternal age and depression is moderated by mode of con-ception, with older women using ART having the greatest risk.

MATERIALS AND METHODSSetting, Sample, and RecruitmentThe study methods have been reported in detail elsewhere (14); in brief, after

gaining institutional ethics approval, stratified sampling was used to ensure

sufficient power to test both age and ARTeffects (and their interaction). Nul-

liparous women in the third trimester of pregnancy who were able to speak

English sufficiently well to complete study materials were approached in

the following settings: women who had conceived with ART at one of seven

ART clinics in two large Australian cities; women who had conceived spon-

taneously (SC) at private and public hospital antenatal clinics; and classes in

the geographic vicinity of the ART clinics. A priori power calculations for the

PATPA project (effect size f2¼ 0.15, a¼ 0.05, power¼ 0.95, number of pre-

dictors ¼ 14) ensured power >.95 for these analyses with a predicted total

sample size of 480.

Women were stratified to maternal age groups (younger, middle, older)

based on their predicted age when their baby was due. There is inconsistency

across studies in the operational definition of the older mother (15). Women

whowould beR37 years at the time of first birth, the age at which pregnancy

rates decrease exponentially (19) and at which previous research has shown

diminishing benefits of delayed parenthood (12), were categorized as older.

Fertility and Sterility�

Gleicher et al. (19) argue that ‘‘the age of 37–38 years, when the female fer-

tility decline accelerates, would appear a more logical point of definition for

the beginning of advanced reproductive age’’ (p. 640). The younger age

group (20 to 30 years old) included those who would be at or younger than

the average age at first birth in Australia (29.9 years of age) (4) when they

gave birth. Women giving birth between the ages of 31 and 36 years

constituted the middle age group. Based on the power analysis, women

were admitted to the study until target numbers (n ¼ 80) were achieved/

exceeded in each maternal age group of spontaneously and ART conceiving

women, respectively. After they had provided informed consent, women

participated in structured telephone interviews and completed validated

self-report questionnaires at approximately 30 weeks of pregnancy and

4 months postpartum. A diagnostic interview for depression was conducted

at 4 months postpartum.

Predictor Variables: Age and Mode of ConceptionThe age groups were as described previously. We adopted the standard def-

inition of ART (20) as involving both oocyte retrieval and embryo transfer.

During the baseline interview, 27 women recruited as SC reported in their

pregnancy interview the use of a fertility treatment (ovulation induction or

artificial insemination) that did not involve oocyte retrieval or embryo

transfer and were therefore excluded. Women recruited from non-ART

hospital sites who then reported conception through ART (through a clinic

other than the ART recruitment sites for the study) were reallocated to the

ART group.

Control variables We included demographic variables (more than or less

than 10 years of schooling; single), fertility and obstetric variables related to

maternal age and/or ART conception (multiple birth, cesarean birth, previous

miscarriage, termination); and psychosocial variables identified in a system-

atic review as primary predictors of postnatal depression (21), namely,

history of depression (here indexed as symptoms during pregnancy), relation-

ship quality, social support, and recent adverse life events. The depression

symptoms were measured during pregnancy (as a control variable) and at

4 months postpartum to enable validation of the diagnostic interview by

use of the 10-item Edinburgh Postnatal Depression Scale, which has also

been validated for use during pregnancy (22).

Relationship quality was assessed using the Intimate Bonds Measure (23)

(IBM), which yields two independent subscales: Care (expression of warmth/

affection) and Control (criticism/restriction of freedom).We used the Control

scale completed 4 months after birth, as this dimension has been most closely

linked to mood in the perinatal period (24). To retain single women in the

multivariate analyses, a categorical variable with three levels was generated

for the control variable: 0 (no partner); 1 (partner low on control, <90th

centile); 2 (partner high on control, R90th centile). The 90th centile was

selected because recent evidence has indicated that scores in this range are

associated with mental health problems (25). Social support was assessed

by a single interview question askingmothers to rate their overall satisfaction

with practical support with housework and infant care from a variety of sour-

ces (e.g., partner, friends, paid help). Scores were rated from 1 to 5, with

higher scores indicating more support. Mothers also answered a single-

item question (yes/no) about adverse life events: ‘‘Are there events in your

life at present, quite apart from having a baby, that are worrying or

distressing?’’

Outcome Variable: Major Depressive Disorder (MDD)Mothers were interviewed (telephone) when their babies were 4 months old

using theMini-Plus International Neuropsychiatric Interview (MINI, version

5.0.0) (26) to assess depression. This structured, diagnostic interview con-

tains questions about the intensity, frequency, and duration of specific symp-

toms, and the degree of distress or impairment associated with them. The

instrument is designed to retrospectively assess lifetime episodes of depres-

sion. In the current study, the usual phrase ‘‘past episode’’ was modified to

specify occurrences in the preceding months ‘‘since the baby was born.’’

The MINI is both reliable and valid and demonstrates good concordance

with both the Structured Clinical Interview for DSM diagnoses (SCID)

(26) and the Composite International Diagnostic Interview for ICD-10

1219

Page 3

(CIDI) (27). In the present study, women who met the diagnostic criteria for

depression scored statistically significantly higher (t (539) ¼ 5.65; P<.001)

on concurrent depression symptoms measured by the Edinburgh Postnatal

Depression Scale (EPDS). Mothers also reported details of any antidepres-

sant medications they had taken in the preceding month.

Data AnalysisTo ensure consistency in regression models, only women with complete data

were included (except for comparisons on partnership variables such as years

married, which were not applicable for single women). Data screening re-

vealed statistically significant skewness (R2.58, P<.001) on the depression

symptom measure and satisfaction with support postnatally. Appropriate

transformations were applied (square root and logarithmic transformation,

respectively).

All analyses are calculated on these transformed scores. However, to allow

cross-study comparisons, descriptive statistics present the untransformed

data in original units of measurement. For all analyses, probability values

P%.05 were considered statistically significant.

Univariate analyses were used to compare age groups on control variables.

A series of logistic regression models were used to test the hypotheses relat-

ing maternal age and a diagnosis of MDD according to the methods of Aiken

and West (28). The first model tested for the main effect of maternal age-

group on MDD rates. The second model tested the main effect of maternal

age-group in the presence of mode of conception (spontaneous or with

ART) and their interaction. The final model examined these associations

while controlling for the primary psychosocial risk factors for MDD as

well as obstetric and fertility correlates of conception at an older age and/

or through ART. As maternal age group and IBM-control were categorical

variables with three levels, a set of two dummy variables was created for

each. The reference category for this set of contrasts was the oldest age group

(%37 years of age). The outcomewas experience of MDD since birth (coded

1). The reference group in the final model for IBM-control was women with-

out a partner. The odds ratios (OR) and 95% confidence interval (CI) are re-

ported. Statistically significant interactions were followed up by recomputing

the logistic regression within each of the conception groups. The models

were repeated using age as a continuous variable and linear regression. Find-

FIGURE 1

Details of sample recruitment, exclusion, retention. Note: Women lost t

group.

Recruitedfrom IVFServices

Recruited from general

/ othermaternityservices

Approached(& eligible)

542 637

Consented 329 (61%) 462 (73%)

Completebaselineinformation

297 295

Lost to contact=3

Lost to contact=13

Completepostnatalinterview data

294 (99%) 282 (96%)

CompletepostnatalInt + Qdata

275 (93%) 266 (90%)

SOME REALLOCATION ACCORDING TO TRUE MOD

McMahon. Older first-time mothers and postpartum depression. Fertil Steril 2011.

1220 McMahon et al. Older first-time mothers and postpart

ings were not markedly different with age as a continuous variable and are

not reported.

RESULTSSee Figure 1 for sampling, recruitment, and retention details. Ofthose who consented to participate, 90% and 64% of women con-ceiving with ART and spontaneously, respectively, completed thebaseline assessment. Complete postpartum data (interview and ques-tionnaire) were available for 93% of the ART participants and 90%of the SC participants who provided baseline data. Data are not avail-able on thosewho declined participation, but demographic data wereavailable for women who completed only some components of eachassessment. Attrition at both time points was more likely for womenin the younger age group and for women with 10 or fewer years ofeducation (P>.05). Pregnancy interviewswere conducted in the thirdtrimester (mean ¼ 31.6, standard deviation [SD] ¼ 2.5 weeks) andpostnatal interviews when babies were approximately 4 monthsold (mean ¼ 19.15 weeks, SD ¼ 3.14 weeks).

Descriptive age-group data and analysis of variance (ANOVA) Fvalues are presented in Table 1. We also note comparisons with Aus-tralian population data when available. Due to our sampling strategy,the use of ARTwas higher (50%) than in the general Australian pop-ulation 4% (4), and the participants were more educated and lesslikely to be single parents (29). The multiple pregnancy rates alsowere somewhat higher (single-embryo transfer is now acceptedpractice in Australia) (4).

The main differences observed between the maternal age groupswere expected as a result of stratified sampling (by age, ART use)and reflected known correlates of conception and childbearing atan older age; for example, greater likelihood of cesarean section:c2 (2) ¼ 21.77 (P¼.001); the greater likelihood of prior miscar-riages: c2 (2) ¼ 14.91 (P¼.001); and terminations as c2 (2) ¼12.41 (P¼.001). The age groups did not differ on education, the

o follow-up were more likely to be in the younger (%30 years) age

TOTAL

1179

791 (67%)

Withdrew consent=37Lost to contact=30 Incomplete baseline data

(completed interviewbut not q’naire)=102

Ineligible (age <20)=3

27Excludedas receivedinfertilitytreatmentother thanART

592

Lost to contact=16

576 (97%)

541 (91%)

E OF CONCEPTION

um depression Vol. 96, No. 5, November 2011

Page 4

TABLE 1Mean and standard deviation of study variables according to maternal age groups (N [ 541).

VariableAustraliannorms

Younger<31 y

(n [ 155)

MiddleR31 and <37 y

(n [ 206)

OlderR37 y

(n [ 180)

(dfs), F valueor chi-square

value

Demographica

Maternal age at birth, mean (SD) 29.9 28.17 (1.7)b 33.44 (1.8)c 39.24 (2.1)d (2, 538) 1,437.98e

%Grade 10 education, % yes (n) 30% 5.2 (8) 3.4 (7) 5.6 (10) (2) 1.16Single/no partner, % yes (n) 20% 1.3 (2)b 2.4 (5)b 5.0 (9)c (2) 4.32

Reproductive factors

Previous medical termination, %

yes (n)

10.3 (16)b 12.1 (25)b 22.8 (41)c (2) 12.41e

Previous miscarriage, % yes (n) 18.7 (29)b 22.8 (47)b 36.1 (65)b (2) 14.91e

Used ART, % yes (n)f 3% 32.3 (50)b 46.1 (95)c 72.2 (130)d (2) 56.18e

Multiple pregnancy, % yes (n) 1.6% 3.2 (5) 3.4 (7) 6.7 (12) (2) 3.17

Cesarean section, % yes (n) 32% 38.7 (60)b 35.4 (73)b 57.8 (104)c 21.77e

Psychosocial risk factors

Pregnancy depression scores,

mean (SD)

6.6 5.75 (4.3) 5.06 (4.3) 4.75 (4.5) (2, 538) 2.86

Satisfaction with practical support,

mean (SD)

— 4.25 (0.9) 4.12 (1.0) 4.04 (1.04) (2, 538) 1.75

Partner R90th centile control,

% (n)g— 13.1 (20) 10.0 (20) 15.8 (27) (2) 2.84

Recent/current stressful life events,

% yes (n)

— 30.3 (47) 36.4 (75) 42.2 (76) (2) 5.09

a Laws and Sullivan, 2009 (4) and Milgrom, et al. 2008 (6).b,c,d Values with no superscripts do not differ significantly. Values with different superscripts are statistically significantly different (P< .01) across age-groups

as follows: c is statistically significantly different from b; d differs from both b and c.e P< .001.f Assisted reproductive technology (ART) use does not reflect population values due to stratified sampling.g Single women excluded.

McMahon. Older first-time mothers and postpartum depression. Fertil Steril 2011.

proportion who were single or having a multiple birth, or on thepsychosocial risk factors associated with vulnerability to postnataldepression (all P>.10).

The ARTand SC groups were not statistically significantly differ-ent on the control variables of being single, previous terminations,satisfaction with partner support, or experience of stressful events(P>.10). A higher proportion of women in the ART group had %10 years education (7.3% vs. 1.9%, c2 (2)¼ 8.29;P¼.002), previousmiscarriage (32.7% vs. 19.2%, c2 (2) ¼ 12.89; P¼.001), multiplepregnancy (6.9% vs. 1.9%, c2 (2) ¼ 8.06; P¼.001), and cesareanbirth (49.1% vs. 38.3%, c2 (2) ¼ 6.34; P¼.007). Women in theART group reported statistically significantly lower depression

TABLE 2Number and percentage ofmothersmeeting criteria formajor depr

reproduction use.

Mode of conceptionYounger<31 y

MR31

SC 11/105 (7.1) 4/1

ART 3/50 (6.0) 11/

Total 14/155 (9.0) 15/2

Note: SC ¼ spontaneous conception; ART ¼ assisted reproduction technique.

McMahon. Older first-time mothers and postpartum depression. Fertil Steril 2011.

Fertility and Sterility�

symptoms during the third trimester of pregnancy: t (539) ¼ 3.41;ART: mean (untransformed) ¼ 4.59, SD ¼ 4.1; and SC: mean (un-transformed) ¼ 5.74, SD ¼ 4.5. However, the mean for both groupswas comparable to normative Australian data (29).

Table 2 presents theMDD rates during the first 4 months after birthaccording to maternal age group and use of ART. The overall rate was7.9% (43 cases among the 541 participants). The incidence ofMDD inthe SC group was lowest in the middle age-group (3.6%, n ¼ 4). Thereverse was true for the group that conceived with ART where thehighest rate was observed in the middle age-group (11.6%, n ¼ 11).Twenty-four women (4%) reported taking antidepressant medicationduring the last month, but there were no age-group (7 young, 10

essive disorder according tomaternal age group and assisted

iddleto <37 y

OlderR37 y Total

11 (3.6) 6/50 (12.0) 21/266 (8.7)

95 (11.6) 8/130 (6.2) 22/275 (7.9)

06 (7.3) 14/180 (7.8) 43/541 (7.9)

1221

Page 5

middle, and 7 older) or mode of conception (13 SC and 11 ART)differences (P>.10; medication data not shown in table). The logisticregressionmodels forMDD rates are presented in Table 3. Odds ratios(OR) and confidence intervals (CI) are reported.

Maternal age group was not statistically significantly related tothe likelihood of MDD diagnosis in any model nor was the modeof conception. The maternal age group by mode of conception inter-action was statistically significant in models 2 and 3 (both P¼.03).Separate follow-up logistic regression analyses were subsequentlyconducted for the SC and ART groups to identify any statisticallysignificant difference in the subgroups. However, none of thesefollow-up contrasts were statistically significant (P>.05). Of thecontrol variables, having only 10 or fewer years of educationwas sta-tistically significant (OR ¼ 3.60; 95% CI, 1.18, 10.98) in predictingMDD rates. Most other control variables were in the expected direc-tion, with the exception of multiple birth and experience of adverselife events, which might be expected to be associated with a greaterlikelihood of depression. We also examined variables specific to SC(unplanned pregnancy) and ART conception (use of donor gametes,number of treatment cycles). Thewithin-group analyses showed thatin the SC group theMDD rates were not statistically significantly as-sociated with the pregnancy being planned or unplanned (adjustedOR¼ 1.61; 95%CI 0.46, 5.70), and in the ART group theMDD rateswere not associated with the use of donated gametes (adjusted OR¼1.55, 95% CI .28, 8.7) or the number of previous failed ART cycles(adjusted OR ¼ 1.01; 95% CI, .87, 1.16).

DISCUSSIONThe results do not indicate an increased prevalence of postpartumdepression during the initial 4 months after a first birth in older

TABLE 3Odds ratios (95% confidence interval) for incremental logistic re

(N [ 541).

Model 1

Predicted maternal age group P¼ .83

%30 y 1.18 (.54, 2.55)

31–36 y .93 (.44, 1.99)

R37 y (reference) —Mode of conceptiona

SC (reference)

ARTMode of conception (by age group)

%30 y

31–36 y

R37 y (reference)Control variables: reference group

%Grade 10 education: yes

Had termination: yes

Had miscarriage: yesTwins: yes

Cesarean: yes

Pregnancy depression scorePractical support score

Recent stressful life events: yes

Relationship controlb

a Mode of conception: 0 ¼ spontaneous conception (SC); 1 ¼ used assisted reb Single women were included as the reference group: 0 ¼ no partner; 1 ¼ part

McMahon. Older first-time mothers and postpartum depression. Fertil Steril 2011.

1222 McMahon et al. Older first-time mothers and postpart

mothers, nor in mothers conceiving through ART irrespective ofage. The MDD rates in older mothers and mothers conceivingwith ART were at the lower end of published norms despiteevidence of greater fertility and obstetric risk in these mothers.Findings suggest, however, that mode of conception may moderateage effects, but the small number in some study cells (young ARTand older SC) and resulting large confidence intervals mean thatthis interaction must be interpreted cautiously. The observed ratesof depression suggest that MDD may be highest when mode ofconception (ART, SC) is less typical for the age group (i.e., womenusing ART in their early 30s, and women conceiving spontane-ously in their late 30s and early 40s had relatively higher depres-sion rates). These suggestive findings warrant further investigationin larger samples.

Strengths and Limitations of the StudyThe methodology in this large prospective study was robust. Thesample included women conceiving with ART and spontaneously,which made it possible to examine for the first time the main effectsof age and mode of conception as well as their interaction. Overall,participants had a high sociodemographic status typical of womenconceiving at an older age (2, 8), of women conceiving throughART (5), and also of Australian mothers conceiving spontaneouslywho volunteer to participate in research. We acknowledge, however,that our strategy of stratified sampling, rather than starting witha large obstetric sample, means that the sociodemographic profilein this study is not typical of women having babies in Australia. Afurther concern regarding this sample profile was the relativelyhigher attrition of mothers in the younger age-group and motherswith lower levels of education.

gression models predicting major depressive disorder rates

Model 2 Model 3

P¼ .11 P¼ .12

.86 (.30, 2.47) .97 (.32, 2.98)

.27 (.07, 1.02) .29 (.07, 1.14)

— —P¼ .20 P¼ .20

— —

.48 (.16, 1.46) .47 (.15, 1.49)P¼ .03 P¼ .03

1.13 (.20, 6.40) 1.35 (.23, 7.89)

7.28 (1.44, 36.89) 8.41 (1.58, 44.81)

— —

3.60 (1.18, 10.98)

1.03 (.42, 2.51)

1.34 (.65, 2.73).96 (.20, 4.60)

1.18 (.60, 2.32)

1.27 (.92, 1.75)2.54 (1.00, 6.40)

.52 (.27, 1.00)

1.12 (.67, 1.87)

productive technology (ART).

ner low on control; 2 ¼ partner high on control.

um depression Vol. 96, No. 5, November 2011

Page 6

The assessment of depression using a structured diagnostic inter-view represents an important advance, as previous research witholder mothers and with mothers conceiving through ART was lim-ited by reliance on self-report symptom measures (5, 17). Thediagnostic interview identifies symptoms likely to be clinicallysignificant and associated with impaired function, whereas use ofsymptom screening tools may result in overestimates of true rates(30, 31), potentially overpathologizing processes of adjustment tomotherhood (31). Finally, the analytic approach controlled forwell-recognized psychosocial (primary) and obstetric risk (second-ary) factors for postnatal depression that may be confounded withthe effects of both maternal age and ART conception.

The main limitation was related to base rates for some cells of thedesign, namely, the smaller sample size per group for older mothersconceiving spontaneously (n ¼ 50) and for younger mothers con-ceiving with ART (n¼ 50). This reflects the relatively low numbersof these two groups in the population. Our total sample size was suf-ficient to detect medium to small age-group effects in depression (ifthese existed), but follow-up contrasts in specific cells producedlarge confidence intervals.

Overall, the rate of depression in this sample (7.9%) was at thelower end of commonly reported Australian community rates of10% to 15% (6), and closer to the rate of 7.1% reported in a recentsystematic review (30). It is noteworthy that even the highest MDDrates (i.e., ART mothers aged 31 to 36 years and mothersR37 yearsof age in the spontaneous conception group,z11%) were relativelylow compared with reported community rates. This is likely to bepartly attributable to the high sociodemographic status of the sam-ple. This high socioeconomic profile is characteristic of oldermothers and of women using ART to conceive, and therefore onecan reasonably conclude that the risk profile for MDD in these pop-ulations is low. It is, of course possible that there was some under-reporting of depression, commonly noted in postnatal women dueto the considerable stigma associated with depression after child-birth (32). Women conceiving through ART, in particular, mayfeel scrutinized and inclined to report in socially desirable ways (5).

CONCLUSIONDemographers agree that the trend toward older age at first birth maybecome the norm (1, 19). Although it is important that women are

Fertility and Sterility�

educated about age-related declines in fertility (2, 33), personalchoice is just one of many complex factors contributing to laterfirst-time parenthood (2, 8). The results of this study indicate thatolder mothers and those conceiving through ART are not atgreater risk of postpartum depression, at least in the early monthsafter birth, which underpins the importance of not stigmatizingand labeling older mothers and avoiding systematic biases inattitudes and approaches to care (2, 3, 19). Indeed, the labeling ofolder mothers as high risk may be a major contributor to maternalangst (34).

The intriguing finding suggesting that older mothers who con-ceive spontaneously may be at higher risk of postpartum depres-sion requires further investigation, as the number of thesewomen in the current study was relatively small. Our findingsalso add to a growing body of research indicating that the psycho-logical adjustment in the perinatal period of women conceivingthrough ART is generally comparable with spontaneously conceiv-ing women (17).

Further longitudinal follow-up observation is important for botholder mothers and those conceiving through ART. The transitionto parenthood is just the beginning of the parenting journey, andnew challenges may emerge, particularly around attempts to havesubsequent children. The onset of menopause when children arequite young may also present challenges (15). There is considerableevidence that life-span vulnerability to depression in women isgreatest during midlife (35), and it remains to be seen how olderfirst-time parents, approaching their 50s, cope with the demandsof caring for young children while potentially also caring for agingparents, managing the physiologic challenges of menopause, andmanaging the demands of the workforce. Most Australian womenhave not returned to the workforce by 4 months postpartum (36),and future studies could also explore how older mothers, who maybe more emotionally committed to their careers (37), manage thecombined demands of paid work and childcare.

Finally, our study focused on clinically significant postnatal de-pression. Research has shown that the subjective experience of thetransition to parenthood is qualitatively different for older mothers(3), and further investigations using a range of methodologies andassessment methods are needed to fully capture the experience ofolder first-time mothers.

REFERENCES

1. De Graaff A, Land J, Kessels A, Evers J. Demo-

graphic age shift toward later conception results in

an increased age in the subfertile population and an

increased demand for medical care. Fertil Steril

2011;95:61–3.

2. Mills TA, Lavender T. Advanced maternal age. Ob-

stet Gynaecol Reprod Med 2010;21:107–11.

3. Carolan M. ‘‘Doing it properly’’: the experience of

first time mothering over 35 years. Health Care

Women Int 2005;26:764–87.

4. Laws P, Sullivan EA. Australia’s mothers and babies

2007. Perinatal statistics series no. 23. Cat. no. PER

48. Canberra: Australian Institute of Health and

Welfare; 2009.

5. Hammarberg K, Fisher JRW,Wynter KH. Psycholog-

ical and social aspects of pregnancy, childbirth and

early parenting after assisted conception: a systematic

review. Hum Reprod Update 2008;14:395–414.

6. Milgrom J, Gemmill A, Bilszta J, Hayes B, Barnett B,

Brooks J, et al. Antenatal risk factors for postnatal de-

pression: a large prospective study. J Affect Disord

2008;108:147–57.

7. Hammarberg K, Fisher JRW, Rowe HJ. Women’s ex-

perience of childbirth and post-natal healthcare after

assisted conception. Hum Reprod 2008;23:1567–73.

8. Carolan, M., Frankowska, D. Advanced maternal age

and adverse perinatal outcome: a review of the evi-

dence. Midwifery. Published online September 29,

2010.

9. Berryman JC, Thorpe KJ, Windridge KC. Age and

parenting. In: Niven C, Walker A, editors. Current is-

sues in infancy and parenthood. Oxford: Butterworth,

Heinemann; 1999. p. 60–88.

10. Hoff E, Laursen B, Tardif T. Socioeconomic status

and parenting. In: Bornstein M, editor. Handbook of

parenting, vol. 2. Mahwah, NJ: Lawrence Erlbaum;

2002. p. 231–53.

11. Casey Jacobs M. The older infertile patient. In:

Burns LH, Covington SN, editors. Infertility counsel-

ing: a comprehensive handbook for clinicians. New

York: Parthenon; 2000. p. 267–83.

12. Mirowsky J, Ross CE. Depression, parenthood and

age at first birth. Soc Sci Med 2002;54:1281–98.

13. McMahon C, Gibson F, Allan J, Saunders DM.

Psychosocial adjustment during pregnancy for older

couples conceiving through assisted reproductive

technology. Hum Reprod 2007;22:1168–74.

14. McMahon CA, Boivin J, Gibson FL, Hammarberg K,

Wynter K, Saunders D, et al. Age at first birth, mode

of conception and psychological wellbeing in preg-

nancy: findings from the Parental Age and Transition

to Parenthood (PATPA) study. Hum Reprod

2011;26:1389–98.

15. Boivin J, Rice F, Hay D, Harold G, Lewis A, van den

Bree M. Associations between maternal older age,

family environment and parent and child wellbeing

in families using assisted reproductive technologies

to conceive. Soc Sci Med 2009;68:1848–55.

16. Hammarberg K, Rowe H, Fisher JRW. Early post-

partum adjustment and admission to parenting ser-

vices in Victoria, Australia after assisted conception.

Hum Reprod 2009;24:2801–9.

17. Ross L, McQueen K, Dennis C, Bloomberg L.

Risk for postpartum depression associated with

1223

Page 7

assisted reproductive technologies and multiple

births: a systematic review. Hum Reprod Update

2011;17:96–106.

18. Fisher JRW, Hammarberg K, Baker HWG. Assisted

conception is a risk factor for perinatal mood distur-

bance and early parenting difficulties. Fertil Steril

2005;84:426–30.

19. Gleicher N, Weghofer A, Barad D. Too old for IVF:

are we discriminating against older women? J Assist

Reprod Genet 2007;24:639–44.

20. Zegers-Hochschild F, Nygren K-G, David

Adamson G, de Mouzon J, Lancaster P, Mansour R,

et al. The ICMART glossary on ART terminology.

International Committee Monitoring Assisted

Reproductive Technologies. Hum Reprod 2006;21:

1968–70.

21. Scottish Intercollegiate Guidelines Network. Postna-

tal depression and puerperal psychosis: a national

clinical guideline. Publication no. 60. Edinburgh:

SIGN; 2002.

22. Cox J, Chapman G, Murray D, Jones P. Validation of

the Edinburgh Postnatal Depression Scale (EPDS) in

non-postnatal women. J Affect Disord 1996;39:

185–9.

1224 McMahon et al. Older first-tim

23. Wilhelm K, Parker G. The development of a measure

of intimate bonds. Psych Med 1988;18:225–34.

24. Fisher J, Feekery C, Rowe-Murray H. Nature, sever-

ity and correlates of psychological distress in women

admitted to a private mother-baby unit. J Paed Child

Health 2002;38:140–5.

25. Fisher J, Tran T, La BT, Kriitmaa K, Rosenthal D,

Tran T. Common perinatal mental disorders in north-

ern Viet Nam: community prevalence and health care

use. Bull World Health Organ 2010;88:737–45.

26. Sheehan DV, Lecrubier Y, Sheehan KH, Janavs J,

Weiller E, Keskiner, et al. The validity of the Mini

International Neuropsychiatric Interview (MINI)

according to the SCID-P and its reliability. Eur

Psychiatry 1997;12:232–41.

27. World Health Organization. Composite International

Diagnostic Interview (CIDI), version 2.1). Geneva:

World Health Organization; 1997.

28. AikenLS,West SG.Multiple regression: testing and in-

terpreting interactions.Newbury Park,CA: Sage; 1992.

29. Beyondblue national postnatal depression program

prevention and early intervention 2001: 2005 final re-

port. Vol. 1, National screening program. Available

at: http://www.beyondblue.org.au.

e mothers and postpartum depression

30. Gavin N, Gaynes B, Lohr K, Meltzer-Brody S,

Gartlener G, Swinson T. Perinatal depression: a sys-

tematic review of prevalence and incidence. Obstet

Gynecol 2005;106:1071–83.

31. Matthey S. Are we overpathologising motherhood? J

Affect Disord 2010;120:263–6.

32. McCarthy M, McMahon C. Acceptance and experi-

ence of treatment for postnatal depression in a com-

munity mental health setting. Health Care Women

Int 2008;29:618–37.

33. Soules MR. The story behind the American Society

for Reproductive Medicine’s prevention of infertility

campaign. Fertil Steril 2003;80:295–9.

34. Mansfield PK. Midlife childbearing: strategies for in-

formed decision making. Psychol Women Q

1988;12:445–60.

35. Wilhelm K, Parker G, Geerligs L, Wedgwood L.

Women and depression: a 30 year learning curve.

Aust NZ J Psychiatry 2008;42:3–12.

36. Australian Bureau of Statistics. Pregnancy and em-

ployment transitions Australia, 2005. Cat. no.

4913.0 Canberra: ABS.

37. McDonaldP.Low fertility inAustralia: evidence, causes

and policy responses. People and Place 2000;8:6–20.

Vol. 96, No. 5, November 2011