- Top of page
Miscarriage is the most common complication of pregnancy, occurring in up to 20% of all clinically recognised pregnancies. Women are often unprepared for the loss and possibly traumatic event, and thus experience a range of psychological reactions, ranging from grief to anxiety and depression.
There is increasing evidence that miscarriage is associated with significant psychological morbidities.[3-8] Friedman and Gath reported that up to 50% of women developed depressive disorder following miscarriage. A local study by Lee et al. found a lower rate (~10%) of Hong Kong Chinese women suffered from major depression 6 weeks after a miscarriage, compared with a 1.4% background prevalence.
The level of psychological impact is believed to be higher in pregnancies conceived following assisted reproduction. It would not be surprising to find that miscarriage results in greater disappointment and related psychological sequelae in this group of women. Indeed, there has been evidence that subfertility leads to psychological distress in excess of community norms. In one study, about one-third of patients about to undergo in vitro fertilization (IVF) treatment had depression. Domar et al. also found significantly higher levels of depressive symptoms in subfertile women, when compared with fertile women, and found mood disorders of a magnitude similar to those experienced by patients with chronic medical conditions. The elevated levels of anxiety and depressive symptoms among women who undergo assisted reproductive technologies was confirmed and found to be persistent over time in a recent study. In addition, the duration of subfertility and need for assisted reproduction are both postulated to be associated with elevated emotional stress after miscarriage.
The current study aims to compare stress and anxiety-depression levels following an early pregnancy loss between women who conceived naturally and those who conceived after assisted reproduction. It is hoped that the results can shed light on the holistic management of these women, and aid in the early detection of significant psychological reactions in avoiding long-term morbidities.
- Top of page
This prospective cohort study was conducted in a university-affiliated, tertiary referral hospital that assists with about 4000 deliveries a year. From April 2008 to October 2010, women who conceived naturally and suffered from miscarriage at less than 12 weeks of gestation (first-trimester miscarriage) were recruited during their attendance in the early pregnancy assessment clinic or upon admission to the general gynaecological ward. Subfertile patients who conceived after treatment at the Centre of Assisted Reproduction, Queen Mary Hospital of the University of Hong Kong, and were diagnosed with first-trimester miscarriage were recruited to our study. Every patient gave an informed written consent prior to participating in the study, which was approved by the Institutional Review Board of the University of Hong Kong/Hospital Authority Hong Kong West Cluster. Patients were excluded from the study if they were non-Chinese, had a known history of psychiatric illness, had a history of recurrent miscarriages (defined by three or more consecutive miscarriages), or already had a living child. Depending on their diagnoses and clinical conditions, patients were managed according to the departmental standard protocols on expectant, medical, or surgical treatment for miscarriage, after informed choices had been made.
Social demographic data and age, social class, education level, type and duration of subfertility, treatment received, and previous obstetric history were collected for each woman. Semi-structured interviews with two standardised questionnaires [the 12-item General Health Questionnaire (GHQ-12) and the 22-item Revised Impact of Event Scale (IES–R)] were conducted by a dedicated research nurse at 1, 4, and 12 weeks after the diagnosis of first-trimester miscarriage.
The GHQ-12, which consists of subscales that measure somatic symptoms, anxiety, insomnia, social dysfunction, and severe depression, is one of the most extensively used tests for the assessment of mental wellbeing, and for detecting psychiatric morbidity, in general medical and outpatient settings. It has also been commonly used in previous studies to assess psychological morbidities in patients with subfertility or miscarriage. A score of more than 15 would be considered significant psychological distress, and scores of 11 or 12 indicate typical response levels to miscarriage.[9, 17-20]
The IES-R was designed to assess current subjective distress for any specific life event and psychological response patterns associated with trauma: intrusion, avoidance, and hyperarousal symptoms.[21, 22] These are the three domains included in the diagnosis of post-traumatic stress disorders, the symptoms of which include: the persistent re-experiencing of the event (intrusion); avoidance of stimuli associated with the trauma and numbing of general responsiveness (avoidance); and exaggerated arousal, irritability, or startled response not present before the event (hyperarousal). The reliability and validity of the Chinese version of both questionnaires have been verified with high internal consistency.[17-27]
Psychological intervention would be offered to women with significant distress, as detected by the above assessment scales, or upon interviews when women expressed disturbed wellbeing.
The prevalence of significant psychiatric morbidity after miscarriages in spontaneous pregnancies was 12%, and the corresponding prevalence following assisted reproduction was assumed to be about three times higher. To give a test of significance of 0.05 and a power of 0.8 (sigmastat; Jandel Scientific, San Rafael, CA, USA), the sample size required would be 56 patients in each arm. About 15% of patients might default the subsequent follow-up, and so 65 patients were needed for each arm.
The sociodemographic and clinical data of both arms were compared. Continuous variables were expressed with mean and standard deviation for normal distribution or with median (range) for skewed distribution. Statistical comparisons were carried out by Student's t-test, chi–square test, Mann–Whitney U-test, Kolmogorov–Smirnov test, and Kruskal–Wallis test, whenever appropriate. Statistical analysis was performed using spss 18.0 (Statistical Program for Social Sciences; SPSS Inc., Chicago, IL, USA). The two-tailed value of P < 0.05 was considered to be statistically significant.
- Top of page
During the study period, a total of 167 women with first-trimester miscarriage were recruited. Eight women (two after assisted reproduction and six after natural conception) declined to join the study. A total of 75 women who conceived after assisted reproduction and 75 women who conceived naturally completed the standard questionnaires and follow-ups. Nine (5.7%) were lost to follow-up (all in the naturally conceived group).
Table 1 shows the demographic and clinical data of the natural conception and assisted reproduction groups. The mean age of the assisted reproduction group was significantly higher than that of the natural conception group. Women in the assisted reproduction group were all married and had a lower incidence of smoking and drinking. Apart from five women who had received intrauterine insemination (IUI), the majority of women (70/75; 93%) in the assisted reproduction group conceived after IVF treatment. The number of previous miscarriages and termination of pregnancies were similar in both groups. The duration of gestation at diagnosis was significantly shorter in the assisted reproduction group (67.2 ± 9.4 days) than in the natural conception group (73.0 ± 13.6 days). Despite the demographic differences, the two groups were found to have similar GHQ-12 and IES-R scores at week 1 of the study. Logistic regression analysis was performed and found none of the factors except previous miscarriage (P = 0.021) was significant in predicting psychological distress. Factors including age, marital status, planned/unplanned pregnancy, and medical/surgical evacuation were not predictive.
Table 1. Patient characteristics
| ||Natural conception (n = 75)||Assisted reproduction (n = 75)|| P |
| Age (years) a || || ||<0.001|
| Marital status (married) ||59 (79%)||75 (100%)||<0.001|
| Duration of marriage (months)a || || ||<0.001|
| Occupation |
|Housewife||17 (22.7%)||15 (20.0%)||0.586|
|Manual worker/sales||15 (20.0%)||11 (14.7%)|
|Clerical||38 (50.7%)||40 (53.3%)|
|Professional||5 (6.7%)||9 (12.0%)|
| Smoking |
|Non-smoker||58 (77.3%)||69 (92.0%)||0.025|
|Smoker||13 (17.3%)||3 (4.0%)|
|Ex-smoker||4 (5.3%)||3 (4.0%)|
| Drinking |
|Non-drinker||64 (85.3%)||72 (96.0%)||0.046|
|Social drinker||11 (14.7%)||3 (4.0%)|
|Gravidityb||1.61 ± 0.82 (1–5)||1.57 ± 0.86 (1–5)||0.601|
| No. of previous miscarriages |
|0||64 (85%)||63 (84%)||1.00|
|1||11 (15%)||12 (16%)|
| No. of previous terminations of pregnancy |
|0||52 (69.3%)||60 (80%)||0.395|
|1||15 (20.0%)||8 (10.7%)|
|2||6 (8.0%)||4 (5.3%)|
|3 or more||2 (2.7%)||3 (4.0%)|
| Pregnancy |
|Planned||58 (77.3%)||75 (100%)||<0.001|
|Unplanned||17 (22.7%)|| |
|Gestation length (days)b||73.0 ± 13.6 (49–126)||67.2 ± 9.4 (51–98)||0.005|
| Current pregnancy |
|Singleton||74 (98.7%)||70 (93.3%)||0.209|
|Twins||1 (1.3%)||5 (6.7%)|
| Treatment modality |
|Expectant||26 (34.7%)||11 (14.7%)||<0.001|
|Medical||16 (21.3%)||55 (73.3%)|
|Surgical||33 (44.0%)||9 (12.0%)|
The results of GHQ-12 in both groups at 1, 4, and 12 weeks after miscarriage are presented in Figure 1. Both groups had highest scores at 1 week after the diagnosis of miscarriage [median 17, range 6–33, interquartile range (IQR) 14–20 in the natural conception group; median 17, range 7–32, IQR 12–22 in the assisted reproduction group; P = 0.829]. The median levels reaching >15 in both groups indicated that there was significant psychological distress associated with first-trimester miscarriage. These scores were noted to decrease with time in either group. By 4 weeks, both groups scored less than the distress level (median 9, range 6–28, IQR 7–11, in the natural conception group; median 12, range 5–36, IQR 9–17, in the assisted reproduction group; P < 0.001). For the natural conception group, a score <12 indicated a return to typical response levels by 4 weeks. In the assisted reproduction group, the median score only returned back to the typical response level by 12 weeks (median 8, range 6–19, IQR 7–10, in the natural conception group; median 11, range 6–29, IQR 9–13, in the assisted reproduction group; P < 0.001). The subsequent difference between the two groups noted could represent the baseline difference in anxiety levels, which was statistically higher in the assisted reproduction group.
The results of IES-R in Figure 2 were similar to that of GHQ–12. Both groups showed the highest score at 1 week after miscarriage (median 16, range 1–44, IQR 7–23, in the natural conception group; median 18, range 3–43, IQR 7–25, in the assisted reproduction group; P = 0.494), indicating an immediate response to the traumatic event, followed by a gradual decrease with time between follow-ups. At all times, the mean scores were higher in the assisted reproduction group, even by 12 weeks. The difference was more marked at 4 weeks (median 4, range 0–32, IQR 2–8, in the natural conception group; median 8, range 0–37, IQR 5–14, in the assisted reproduction group; P < 0.001) than at 12 weeks (median 4, range 0–29, IQR 1–7, in the natural conception group; median 5, range 0–32, IQR 3–7, in the assisted reproduction group; P = 0.075). Again, the natural conception group appeared to have a shorter time to return to baseline.
In the subgroup analysis of IES-R, all subscales showed decreasing scores over time (Figure 3; Table 2). Again there were statistically significant differences between the natural conception group and the assisted reproduction group at 4 weeks after miscarriage: avoidance symptoms (median 0.125, range 0.00–1.50, IQR 0.00–0.375; median 0.500, range 0.00–1.88, IQR 0.125–0.625; P < 0.001); intrusion symptoms (median 0.125, range 0.00–1.38, IQR 0.125–0.375; median 0.375, R 0.00–2.00, IQR 0.25–0.625; P < 0.001); and hyperarousal symptoms (median 0.167, range 0.00–1.67, IQR 0.167–0.500; median 0.500, range 0.00–1.50, IQR 0.167–0.833; P = 0.002). Significant differences in intrusion (median 0.125, range 0.00–1.63, IQR 0.000–0.250; median 0.125, range 0.00–1.50, IQR 0.125–0.375; P = 0.044) and hyperarousal (median 0.167, range 0.00–1.00, IQR 0.167–0.500; median 0.333, range 0.00–1.33, IQR 0.167–0.500; P = 0.035) symptoms were still observed at 12 weeks, but not for avoidance symptoms (median 0.125, range 0.00–1.25, IQR 0.000–0.375; median 0.125, range 0.00–1.75, IQR 0.000–0.250; P = 0.264).
Table 2. Subscale breakdown of the IES-R
| || ||Natural conception||Assisted reproduction|| P |
|Week 1||Avoidance||0.625 (0.00–2.13; 0.250–1.125)||0.750 (0.00–2.38; 0.250–1.125)||0.318|
|Intrusion||0.625 (0.00–2.38; 0.375–1.125)||0.750 (0.00–2.38; 0.350–1.125)||0.644|
|Hyperarousal||0.833 (0.00–1.83;0.333–1.167)||0.833 (0.00–2.00; 0.333–1.333)||0.800|
|Week 4||Avoidance||0.125 (0.00–1.50; 0.000–0.375)||0.500 (0.00–1.88; 0.125–0.625)||<0.001|
|Intrusion||0.125 (0.00–1.38; 0.125–0.375)||0.375 (0.00–2.00; 0.125–0.625)||<0.001|
|Hyperarousal||0.167 (0.00–1.67; 0.167–0.500)||0.500 (0.00–1.50; 0.167–0.833)||0.002|
|Week 12||Avoidance||0.125 (0.00–1.25; 0.000–0.375)||0.125 (0.00–1.75; 0.000–0.250)||0.264|
|Intrusion||0.125 (0.00–1.63; 0.000–0.250)||0.125 (0.00–1.50; 0.125–0.375)||0.044|
|Hyperarousal||0.167 (0.00–1.00; 0.167–0.500)||0.333 (0.00–1.33; 0.167–0.500)||0.035|
The modality of treatment to evacuate the uterus differed between the natural conception group and the assisted reproduction groups. Most women from the assisted reproduction group (55, 73.3% versus 16, 21.3%) opted for medical treatment compared with surgical evacuation (9, 12% versus 33, 44%).
During the follow-up period, a total of 11 women (4.7%) were noted to have significant psychological disturbance by subjective reports or objective measures. They were referred for further psychological counselling (Table 3). Seven of them had conceived after assisted reproduction (9.3%), and four had conceived naturally (5.3%). The majority of women were referred after the 4–weeks follow-up, which coincided with the time of the most significant difference in symptoms between the groups.
Table 3. Patients referred for further psychological counselling
| ||Natural conception||Assisted reproduction|
- Top of page
To the best of our knowledge, the present study is the first to directly compare the psychological impact after first-trimester miscarriage between women who conceived naturally and women who conceived after assisted reproduction. Not unexpectedly, women who conceived after assisted reproduction indeed experienced higher stress and anxiety levels after miscarriage when compared with those who conceived naturally, and the results have been confirmed by using the validated GHQ-12 and IES-R questionnaires. Weiss and Marmar suggested that the hyperarousal subscale in the IES-R had good predictive validity with regard to trauma experienced in a life event.[21, 22] The significant difference in the hyperarousal subscale in our study implied that miscarriage transferred a greater psychological trauma to women who conceived after assisted reproduction than those who conceived naturally.
The stress levels experienced by the two groups as indicated by these scores decreased over time, which was consistent with prior knowledge that psychological impact following miscarriage wears off with time. The difference was most marked at 4 weeks, and was still present at 12 weeks (at the end of our observation period). Such a finding might suggest a potential window at the 4-week follow-up to offer therapeutic intervention for adverse psychological sequelae after miscarriage.
There are several strengths to the study. Recruitment, completion of the questionnaire, and follow-up interviews were carried out by a single designated research nurse. All subjects were followed-up by telephone interviews if they could not attend in person, as scheduled. Psychological intervention was offered to women who were at risk of developing long-term morbidity. All women were subjected to standardised questionnaires, and to a certain extent, these interviews may have a positive impact on enhancing their psychological wellbeing after the pregnancy loss.
As mentioned earlier, the heightened response at the 4–week interview suggests that this could create an opportunity for clinicians to involve women in a discussion of any need for intervention, as most women would have a follow-up review after miscarriage around this time. Early identification and appropriate management will aid in reducing long-term morbidities, and will improve the psychological wellbeing of women who have miscarried. This, in turn, will possibly help in preparing them for another pregnancy or treatment cycle. In fact, a recent qualitative, phenomenological study indicated that experiences of pregnancy losses were embedded into women's experiences of subfertility and medical treatment. Marked ambivalence was noted regarding future reproductive options after pregnancy, reflecting prominent feelings of anxiety and grief. It would not be surprising to find patients dropping out of treatment of subfertility prematurely for psychological reasons, which is often underestimated in clinical settings. The significant level of psychological distress among women with subfertility undergoing assisted reproduction suggests that an evaluation of depression and the psychological wellbeing of these women at follow-up visits after miscarriage is essential.
Limitations in the current study are acknowledged. Firstly, there were intrinsic differences between the two groups of women in terms of age, marital status, and duration of pregnancy. It is our local governmental policy that assisted reproduction treatment is to be offered to legally married couples only. Women in the assisted reproduction group generally experienced a longer duration of subfertility before they conceived, which resulted in an apparently higher age and longer duration of marriage in this group. By protocol, women in the assisted reproduction group were arranged to have follow-up ultrasounds for viability after successful treatment. The diagnosis of miscarriage was usually made earlier in this group of women.
Secondly, the duration of the follow-up period was not adequate in this study to show a long-term impact. Previous studies showed that the intensity of psychological responses became less intense by about 6 months, and returned to background levels by 1 year.[30, 31] A recently published local study by Lok et al. also confirmed statistically significant psychological morbidity shortly after miscarriage, with the distress level reducing over time until it was comparable with controls by 1 year. A longer duration for follow-up and longitudinal observation would be helpful to gain a better understanding of the extent of the stress response, and draw a better conclusion for the difference in subsequent recovery period. The duration of the follow-up of the present study was 12 weeks only. This was largely limited by the fact that many patients, especially subfertilie patients in the assisted reproduction group, were very eager to become pregnant and started treatment cycles again after 3 months. Most would have started treatment again by 6–12 months if no other complications arose. It was anticipated in the planning stage that many recruited patients might default the follow-up if the follow-up duration was extended beyond 12 weeks, and the results may differ if patients were pregnant during the follow-up period.
Similar logistic difficulties limited the baseline assessment of psychological morbidities in the two groups in the current study. The scores from two questionnaires at week 1 were taken as the baseline stress levels experienced by the two groups of women. Although the scores were apparently similar, they could have already been altered by the immediate disturbance resulting from the diagnosis of miscarriage. A larger study could be launched to evaluate the true baseline stress and anxiety amongst reproductive women who were keen to get pregnant, including the group awaiting assisted reproduction treatment. It would be of interest to note any difference between the two groups.
It could be understood that women who conceived after assisted reproduction had a preference towards less invasive intervention and the avoidance of surgical complications. Nonetheless, the psychological impact as well as patient satisfaction did not appear to differ between medical treatment and surgical evacuation groups, as reported in a local randomised controlled trial by Lee et al.
Our study was limited to Chinese women, and cultural differences are acknowledged. As the rate of depression after miscarriage was reported to be relatively low among Hong Kong Chinese women, the actual effect of psychological impact after miscarriage after assisted reproduction might be even greater in other regions.
Regarding the psychological intervention, it was purely voluntary and on an optional basis in our setting. There lacked any specific assessment of the pathological responses of the women. Moreover, the two questionnaires are not diagnostic of any psychiatric condition. It is possible that some of the excessive psychological symptoms and morbidities could be overlooked.
Miscarriage is recognised as a traumatic event for women of reproductive age. Qualitative investigations have found that, despite the recognised negative impact on psychological wellbeing, women with miscarriage also experience dissatisfaction with service provision. A lack of understanding and effective communication would undoubtedly impair the ability to detect post-miscarriage morbidity. During the care of women pursuing assisted reproduction, health care providers should be reminded to pay attention to the negative impact on psychological wellbeing and quality of life resulting from unsuccessful treatments. We therefore suggest incorporating psychological assessment into the usual medical follow-up after miscarriage. Bereavement and grief issues should be considered and further explored when these women return. Referral to psychological counselling or group support may be beneficial to these women at an appropriate juncture. The timing of such interventions should be targeted at the most vulnerable period, around 4 weeks after the diagnosis of miscarriage.