Psychosocial stress and treatment outcome following assisted reproductive technology

K.A. Sanders and N.W. Bruce1

Department of Anatomy and Human Biology, The University of Western Australia, Nedlands, W.A. 6907, Australia


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
This study investigated the association between psychosocial stress and outcome of in-vitro fertilization and gamete intra-Fallopian transfer treatment. Ninety women, enrolled for treatment at a private infertility clinic, completed two self-administered psychometric tests (Bi-polar Profile of Mood States, POMS; and State–Trait Anxiety Inventory, STAI) and a questionnaire to ascertain demographic and lifestyle characteristics before the start of treatment. Approximately 12 months later an outcome measure was determined for each participant in terms of whether she was pregnant or not pregnant and the number of treatment cycles undertaken to achieve clinical pregnancy. The women's scores on the psychological tests were similar to published normative scores. On univariate analysis, history of a previous pregnancy was positively related to the probability of pregnancy and full-time employment, a more `hostile' mood state and higher trait anxiety were associated with a lower cumulative pregnancy rate. A Cox multiple regression model found previous pregnancy history, trait anxiety, and the POMS agreeable–hostile and elated–depressed scales to be the most important lifestyle and stress variables predictive of pregnancy. The results emphasize the importance of psychosocial stress in treatment outcome but indicate that the relationships are complex. Further studies are required to validate whether these findings can be generalized to other populations.

Key words: infertility/IVF/prospective study/psychosocial stress


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Many factors may influence the outcome of assisted reproductive technology. These include patient factors such as age, infertility category, duration of infertility, and response to ovarian stimulation (Check et al., 1994Go; Hull, 1994Go; Kenny, 1995Go), as well as patient management practices such as the stimulation protocol used and the number and quality of ova or embryos transferred (Shulman et al., 1993Go; Wang et al., 1993Go; Kenny, 1995Go). These are all parameters routinely measured within the clinic. While previous attempts at relating individual variables to pregnancy have provided some insight into their impact (Nelson et al., 1993Go), frequently there is no adequate reason for failure of a particular cycle. The present study investigated the influence of psychosocial stress on the outcome of in-vitro fertilization (IVF) and gamete intra-Fallopian transfer (GIFT) treatment.

Psychosocial stressors originate in the social relationships or arrangements between individuals and are mediated through psychological processes. For couples involved in IVF and related programmes there are numerous psychosocial stressors to cope with, including: their infertility, the inconvenience of daily injections and blood samples, the perception of low success rates, the wait for results, and financial pressures. These are on top of the psychosocial stressors associated with everyday life. It has been suggested that differences in the ability to cope with stress could have implications for conception rates after treatment (Demyttenaere et al., 1992Go; Merari et al., 1992Go) although there have been relatively few studies in this area.

An individual's ability to cope with stress is conditioned by their past experience and also by their current social environment (Lazarus, 1975Go). Thus, lifestyle factors, such as previous reproductive history, socio-economic status and work commitments, could influence the ability of women to cope with infertility treatment, and the relationship of these factors to treatment outcome is largely unknown. Personality characteristics, such as proneness to anxiety or depression, also determine one's ability to cope with stress (Spielberger et al., 1970Go) and a number of studies have already suggested an association between depression and less favourable treatment outcome (Demyttenaere et al., 1992Go; Thiering et al., 1993Go). Therefore the purpose of the present study was to assess the role of anxiety proneness (trait anxiety), mood states, and specific lifestyle factors on outcome of IVF and GIFT treatment.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Subjects and study design
All new patients of one doctor referred for IVF or GIFT treatment to Concept Fertility Centre, King Edward Memorial Hospital for Women, Subiaco, Western Australia, during the period February 1990 to February 1993, were invited to participate in the study by the first author (K.A.S.) at a patient meeting, ~1–3 months prior to treatment. After receiving a detailed explanation of the study, the women were asked to complete a self-administered questionnaire at home and return it to the investigators prior to the start of their treatment cycle. Participation in the study was on a voluntary basis and each woman was assured complete confidentiality. Due to ethical constraints, it was not possible to follow up women who did not return the questionnaire. The study was approved by the University of Western Australia's Committee for Human Rights.

The questionnaire sought information on general demographic characteristics, aspects of lifestyle and psychological status, and included two standard psychometric tests; the State–Trait Anxiety Inventory (STAI; Spielberger et al., 1970Go) and the Bi-polar Profile of Mood States (POMS; McNair and Lorr, 1982Go). The STAI is comprised of two separate scales; the State scale which measures the degree of anxiety at a particular given time and the Trait scale which assesses the tendency of an individual to respond to stressful circumstances with raised anxiety. The POMS measures six bipolar subjective mood states; one pole measures the positive or desirable aspects of the dimension while the other pole refers to the more negative and less desirable aspects. The six POMS scales are: composed–anxious, agreeable–hostile, elated–depressed, confident–unsure, energetic–tired and clearheaded–confused.

Details of reproductive characteristics were abstracted from records held within the clinic. Approximately 12 months after the entry of the last participant an outcome measure was determined for each participant in terms of whether they were pregnant or not pregnant and the number of treatment cycles undertaken to achieve pregnancy. Since IVF and GIFT cycles were sometimes interchanged and occasionally punctuated by frozen embryo transfer cycles, all treatments commenced were included when calculating the number of cycles to pregnancy. Some women withdrew from treatment or transferred to other clinics before pregnancy. In these cases, the number of cycles was censored at the time of withdrawal; that is the women were not pregnant at that time beyond which their status was unknown. For all analyses, a successful pregnancy was defined as a positive fetal heart beat detected by ultrasound at 8 weeks post oocyte collection.

Statistical analysis
The cumulative probability of pregnancy after successive treatments was estimated by product-limit analysis which takes into account the presence of censored observations. The separate effects of different variables (each one transformed into a two-level factor) on the cumulative pregnancy rate, were tested with Wilcoxon and log-rank tests for homogeneity. The Wilcoxon test is more sensitive to differences across the strata at earlier time periods than that of the log-rank statistic which lends more weight to late time differences. Eighteen potentially prognostic fecundity, lifestyle and psychological variables were investigated. These were previous pregnancy history, age, body mass index, education, work status, socio-economic status, smoking habit, and level of tea, coffee and alcohol consumption, and also the psychological test scores from the POMS (six items) and STAI (two items).

Cox's proportional hazards regression was used to determine the best predictors of time to pregnancy for the sample. In constructing the multivariate model the influence of fecundity (previous pregnancy, age) and lifestyle (work status, education, socio-economic status, smoking habit, tea, coffee and alcohol consumption), variables on time to pregnancy were first examined. These variables were entered into a model and those not contributing to the fit of the model were backwards eliminated. Previous pregnancy history remained as the only significant predictor of time to pregnancy. Previous pregnancy history was then included in a second model along with the psychological test scores. The psychological test scores were entered into the model as continuous variables and those that did not contribute significantly to the fit of the model were backwards eliminated. Hazard ratios (HR) and 95% confidence intervals (CI) were calculated from the Cox regression equation for each variable included in the final model. The analyses were performed using the PHREG procedure of SAS (Release 6.04; SAS Institute Inc., Cary, NC, USA).


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
General
A total of 101 patients consented to participate in the project (45.7% of those invited). Subsequently, five women were excluded from the study because they did not proceed with treatment, another six were excluded because they failed to return the questionnaire before commencing treatment, and one because treatment data were missing. Thus, the analyses were based on 90 women. Sixty-nine women (76.7%) were first-time participants in IVF or GIFT treatment and 21 women (23.3%) had undergone one or more previous treatment cycles. All the study participants were either married or had been living in a de-facto relationship for 5 or more years, a legal requirement for assisted reproductive technology treatment in the state of Western Australia. The age of the women in the study reflected that of the clinic population as a whole and ranged from 23 to 43 years with a mean (± SD) age of 32.6 ± 4.4 years. Subjects' infertility category also reflected that of the clinic population and fell into one of the following broad categories: male factor only (15.5%), female factor only (57.8%), combined male/female (5.6%), and idiopathic (21.1%). The pregnancy rate per treatment cycle commenced for women in the study was 10.9% which was similar to that reported for the State of Western Australia in 1991 and 1992 (Western Australian Reproductive Technology Council Annual Report, July 1993).

Table IGo shows the mean pre-treatment scores for the STAI and the POMS, separately for first-time IVF participants and repeat treatment women. For comparison, published normative scores are also given. Mean psychological scores on all measures were not significantly different between new patients and women who had undergone previous treatment. Mean scores were also comparable to published population norms except for the POMS agreeable–hostile scale where the study group reported significantly less favourable scores (P < 0.025, two-sample t-test).


View this table:
[in this window]
[in a new window]
 
Table I. Mean scores for the State–Trait Anxiety Inventory (STAI) and Bi-polar Profile of Mood States test (POMS)
 
Predictors of outcome
Fifty-six couples participated in more than one treatment cycle at the clinic during the time frame of the study; the median number of treatments undertaken was 2 and the maximum 12. Treatment resulted in a successful pregnancy in 32 women.

Product-limit statistics were used to investigate the separate effects of different stress and lifestyle variables on the number of treatments to success and the results are summarized in Table IIGo. Taken alone, previous pregnancy history, work status, the POMS agreeable–hostile scale and trait anxiety were significantly associated with the number of treatment cycles to pregnancy. Twenty-five women had had a previous pregnancy in the current relationship, and after five treatment cycles the cumulative probability of pregnancy for this group of women was 77.8% compared with 19.6% for women without a pregnancy (Figure 1Go). The probability of pregnancy was lower for the first five treatment cycles for women in full-time paid employment compared with women working part-time or employed in home duties (Figure 2Go). Figure 3Go shows women that who scored towards the agreeable pole of the POMS agreeable–hostile scale experienced a higher success rate than more hostile women across all treatment cycles. For trait anxiety, characteristically anxious women experienced a lower probability of pregnancy during later treatment cycles than non-anxious women (Figure 4Go). No other variables were significantly associated with time to pregnancy.


View this table:
[in this window]
[in a new window]
 
Table II. Influence of single (dichotomous) factors on the number of treatment cycles to pregnancy
 


View larger version (16K):
[in this window]
[in a new window]
 
Figure 1. Comparison of cumulative conception curves for previously pregnant women (dashed line) and women without a previous pregnancy (solid line) calculated with product-limit analysis. The two strata are significantly different (Wilcoxon statistic P = 0.0006, log-rank statistic P = 0.0001).

 


View larger version (16K):
[in this window]
[in a new window]
 
Figure 2. Comparison of cumulative conception curves for women who work full-time (dashed line) and women who do not (solid line) estimated with product-limit analysis. The two strata are significantly different at early time periods (Wilcoxon statistic P = 0.043, log-rank statistic P = 0.230).

 


View larger version (17K):
[in this window]
[in a new window]
 
Figure 3. Comparison of cumulative conception curves for women with low (dashed line) and high (solid line) scores on the Bi-polar Profile of Mood States (POMS) agreeable–hostile scale estimated by product-limit analysis. The two strata are significantly different (Wilcoxon statistic P = 0.003, log-rank statistic P = 0.001).

 


View larger version (16K):
[in this window]
[in a new window]
 
Figure 4. Comparison of cumulative conception curves for women with high (dashed line) and low (solid line) trait anxiety estimated with product-limit analysis. The two strata are significantly different at later time periods (Wilcoxon statistic P = 0.109, log-rank statistic P = 0.037).

 
Cox's regression was used to ascertain the most important lifestyle and psychological variables contributing to the number of cycles required to conceive. Table IIIGo presents the results of two models; model 1 containing only previous pregnancy history and model 2 containing previous pregnancy history, trait anxiety, the squared function of trait anxiety, the POMS agreeable–hostile score and the POMS elated–depressed score. Women with at least one previous pregnancy had a greatly increased probability of conception (unadjusted HR = 5.22, 95% CI 2.24, 12.17) and the magnitude of this effect was somewhat greater after adjustment for the other variables (adjusted HR = 9.65, 95% CI 3.56, 26.9). Lower trait anxiety was associated with a greater risk of pregnancy. Thus, a score of 28 (10th percentile) was associated with an estimated 3.9-fold elevation of risk of pregnancy compared to score 37 (median). However, the relationship was quadratic and women in the highest quartile for anxiety also experienced an increased risk of pregnancy. Lower scores on the POMS agreeable–hostile scale, indicating greater hostility, were associated with a decreased risk of pregnancy. Thus compared to the median score of 26, a score of 17 (10th percentile) was associated with an estimated 80% reduction in risk of pregnancy. In contrast, scores towards the depressive pole of the POMS elated–depressed scale were associated with an increased chance of pregnancy per cycle.


View this table:
[in this window]
[in a new window]
 
Table III. Characteristics of IVF group that are associated with the cumulative probability of pregnancy. Cox's proportional hazard regression models
 

    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
IVF and other assisted reproductive technologies offer great hope to couples in overcoming their fertility difficulties. However, by definition, such techniques are invasive which, over the course of treatment, can engender high levels of psychosocial stress amongst patients (Callan and Hennessey, 1988Go; Merari et al., 1992Go; Boivin and Takefman, 1995Go; Boivin et al., 1998Go). In recent years there has been increased interest in the possible influence of psychosocial stress factors on outcome of assisted reproductive procedures (Demyttenaere et al., 1992Go; Merari et al., 1992Go; Thiering et al., 1993Go; Harlow et al., 1996Go; Stoleru et al., 1997Go; Milad et al., 1998Go). The objective of the present study was to determine the importance of psychosocial stress as a factor in the success rates of IVF and related procedures and, to this end, whether simple measures of personality, mood state and lifestyle have prognostic value for reproductive success. The data demonstrated a number of interesting associations. In particular, full-time employment, trait anxiety and the bi-polar mood states of agreeableness–hostility and elation–depression all related to treatment outcome.

Facchinetti et al. recently reported a reduced pregnancy rate after one treatment cycle for women who worked `outside the home' (Facchinetti et al., 1997Go). In the present study, women in full-time employment had a reduced pregnancy rate for their first five treatment cycles, and these findings are supportive of a role for stress in treatment outcome. It is possible that full-time employment, or working `outside the home', while undertaking IVF treatment, creates a number of additional stressors that patients must cope with. These could include such things as difficulty in arranging leave from work, financial considerations, added time pressure, worry or embarrassment over work colleagues' knowledge of personal issues, and added discomfort due to the stimulatory drugs. It should be noted, however, that work status failed to significantly predict outcome in the multivariate Cox analysis and thus this finding should be interpreted with caution.

Also supportive of a stress–outcome relationship was the finding, in both univariate and multivariate analyses, that a hostile mood was associated with a decreased probability of pregnancy after successive treatment cycles. The question arises whether more `hostile' women are physiologically less responsive to treatment or whether treatment itself is compromised by hostile attitude. It is interesting to note that the clinic provides an information sheet to prospective patients that states: `your co-operation in all aspects (of treatment) mentioned is vital to the success and smooth running of the programme.' Hence, it is possible that the association between hostile mood and decreased chance of pregnancy may have been due to subtle differences in timing or management of routine clinic events. Regardless of actual mechanisms, however, this finding could have important implications for patient counselling and management.

Trait anxiety refers to the tendency of individuals to respond to stressful situations with elevated anxiety (Spielberger et al., 1970Go) and thus, measurement of trait anxiety can be considered an index of ability to cope with stressors. We postulated that highly trait anxious women would be less likely to achieve pregnancy through assisted reproductive technology because they would be less able to cope with the stressors of treatment. The results showed that, at least initially, higher trait anxiety was associated with reduced risk of pregnancy. However, the multivariate regression analysis revealed that the risk of pregnancy actually increased for trait anxiety scores in the highest quartile and this finding is more difficult to interpret.

Few previous studies have addressed the role of trait anxiety in treatment outcome. Demyttenaere et al. found a positive correlation between trait anxiety level and number of treatment cycles amongst women conceiving in a donor insemination programme (Demyttenaere et al., 1988Go). However, they did not take into account women who failed to achieve pregnancy. Merari et al. reported mean trait anxiety, assessed in the early follicular phase, to be similar for unsuccessful women and women who conceived following one IVF cycle (Merari et al., 1992Go), but the levels of anxiety in that study were higher than for normal women and for the women in our study. A recent study reported a trend towards lower trait anxiety in women who became pregnant after one treatment cycle but this was not significant on univariate analysis and multivariate analysis was not applied (Harlow et al., 1996Go). Our study is the first to include multivariate analysis in a prospective study of trait anxiety and fertility treatment outcome after several treatment cycles.

The level of trait anxiety reported by women in this study was within the normal range and consistent with the findings of a number of other studies (Brinsmead et al., 1986Go; Shatford et al., 1988Go; Thiering et al., 1993Go; Boivin and Takefman, 1995Go). An examination of other lifestyle and treatment characteristics for the very highly anxious women failed to reveal any major differences that could account for the increased pregnancy rate observed. According to Spielberger and Rickman highly trait anxious women are more likely to use the psychological defence mechanisms of repression and denial to minimize state anxiety (Spielberger and Rickman, 1990Go). Thus, it could be postulated that these women have a different psychoneuroendocrine stress response pattern that positively influenced treatment outcome. However, further research is required to explore this unexpected but interesting finding.

As well as trait anxiety, state (transient) anxiety was measured in both the POMS and STAI scales. However, in contrast to trait anxiety, neither state scale appeared to have any association with pregnancy rates. This is not an unexpected finding since the women completed the scales after acceptance into the programme and prior to any treatment. Other workers have shown that this time may not demonstrate the anxieties of treatment, but rather, reflect feelings of relief regarding the decision to undertake IVF and general high expectations of success (Callan and Hennessey, 1988Go; Thiering et al., 1993Go). The normal scores found for most other mood states measured in the present study would tend to support this notion.

Clinically depressed women have previously been reported to have lower pregnancy rates for the first six to seven IVF treatment cycles (Thiering et al., 1993Go). In our study, however, the pregnancy rate was increased for women scoring towards the depressed pole on the POMS elated–depressed scale. It must be noted, however, that the POMS scale used in our study was not designed to distinguish clinical conditions. Furthermore, the scale is generally regarded as a measure of transient mood state and, as such, it may be altered by a subsequent change in psychosocial environment. A plausible explanation for the present results, therefore, is that at the time of completing the questionnaire, before entry into the treatment programme, some women were excessively `elated' at the prospect of resolving their fertility problems through IVF. As indicated above, many new IVF participants frequently enter the programme with unrealistic expectations of success (Callan and Hennessey, 1988Go; Reading, 1989Go; Slade et al., 1997Go). However, failure of treatment can lead to severe depression (Baram et al., 1988Go; Dennerstein and Morse, 1988Go; Slade et al., 1997Go). In our study, ~77% of women were new participants and so a high level of `elatedness' might have been expected. For these women, the actual demands of the treatment cycle may have had a major negative impact on their subsequent emotional state, and their chance of pregnancy. It is noteworthy that in the Thiering et al. study (Thiering et al., 1993Go), the association between depression and treatment outcome was less predictive for new participants than for repeat treatment women.

Our finding of a higher pregnancy rate for women with a previous pregnancy in their current relationship confirms that of Haan et al. who demonstrated that the results of treatment are dependent on the couple's previous history rather than on the female partner's history alone (Haan et al., 1991Go). Recently, Molloy et al. found the cumulative pregnancy rates for GIFT and IVF to be higher for the first three treatment cycles, in women who had previously had pregnancies through assisted reproductive technologies (Molloy et al., 1995Go). These authors suggest that the association between previous pregnancy and treatment outcome may relate to differences in oocyte/embryo quality. Unfortunately this factor could not be assessed in our study.

In conclusion, the present study demonstrates an important link between lifestyle and personality characteristics, ascertained before treatment, and the outcome of IVF and GIFT procedures. The findings that full-time work and more hostile mood states are associated with reduced pregnancy rates, in the study group, conform to the original hypothesis that psychosocial stress reduces successful treatment outcome. The findings that trait anxiety and depression are also related to treatment outcome further emphasizes the importance of psychosocial factors but indicates that these relationships are complex. But whatever the actual direction of psychosocial stress and treatment outcome relationships and their underlying mechanisms of action, the present study could have important implications for overall patient management. Although the limited sample size of the present study reduces the ability to generalize the results to other assisted reproductive treatment clinic populations, if these findings are confirmed in other, similar studies, the role of psychosocial support programmes in infertility treatment needs to be more actively investigated.


    Acknowledgments
 
We are grateful to Dr Bruce Bellinge, and the staff of Concept Fertility Centre for allowing us access to their patients and assisting in the collection of the data. We also wish to thank the participating women for giving up their time to complete the questionnaires. The late Dr A.Landauer provided advice on the use of the psychological questionnaires.


    Notes
 
1 To whom correspondence should be addressed Back


    References
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Baram, D., Tourtelot, E., Muechler, E. et al. (1988) Psychosocial adjustment following unsuccessful in vitro fertilisation. J.Psychosom.Obstet.Gynaecol., 9, 181–190.[ISI]

Boivin, J. and Takefman, J.E. (1995) Stress level across stages of in vitro fertilization in subsequently pregnant and nonpregnant women. Fertil.Steril., 64, 802–810.[ISI][Medline]

Boivin, J., Andersson, L., Skoog-Svanberg, A. et al. (1998) Psychological reactions during in-vitro fertilization: similar response pattern in husbands and wives. Hum.Reprod., 13, 3262–3267.[Abstract]

Brinsmead, M., Guttmann, S., Oliver, M. et al. (1986) Demographic and personality characteristics of couples undergoing in vitro fertilisation. Clin.Reprod.Fertil., 4, 373–381.

Callan, V.J. and Hennessey, J.F. (1988) Emotional aspects and support in in vitro fertilization and embryo transfer programs. J.Vitro.Fertil.Embryo Transfer, 5, 290–295.

Check, J.H., Lurie, D., Callan, C. et al. (1994) Comparison of the cumulative probability of pregnancy after in vitro fertilization–embryo transfer by infertility factor and age. Fertil.Steril., 61, 257–261.[ISI][Medline]

Demyttenaere, K., Nijs, P., Steeno, O. et al. (1988) Anxiety and conception rates in donor insemination. J.Psychosom.Obstet.Gynaecol., 8, 175–181.[ISI]

Demyttenaere, K., Nijs, P., Evers-Kiebooms, G. et al. (1992) Coping and the ineffectiveness of coping influence the outcome of in vitro fertilization through stress responses. Psychoneuroendocrinology, 17, 655–665.[ISI][Medline]

Dennerstein, L. and Morse, C. (1988) A review of psychological and social aspects of in vitro fertilisation. J.Psychosom.Obstet.Gynaecol., 9, 159–170.[ISI]

Facchinetti, F., Matteo, M.L., Artini, G.P. et al. (1997) An increased vulnerability to stress is associated with a poor outcome of in vitro fertilization–embryo transfer treatment. Fertil.Steril., 67, 309–314.[ISI][Medline]

Haan, G., Bernardus, R.E., Hollanders, J.M.G. et al. (1991) Results of IVF from a prospective multicentre study. Hum.Reprod., 6, 805–810.[Abstract]

Harlow, C.R. Fahy, U.M. Talbot, W.M. et al. (1996) Stress and stress-related hormones during in-vitro fertilization treatment. Hum.Reprod., 11, 274–279.[Abstract]

Hull, M.G.R. (1994) Effectiveness of infertility treatments: choice and comparative analysis. Int.J.Gynecol.Obstet., 47, 99–108.

Hyndman, J. (1995) Socio-economic Indices for Areas 1981, 1986 and 1991. Department of Public Health, The University of Western Australia, Perth, Western Australia.

Kenny, D.T. (1995) In vitro fertilisation and gamete intrafallopian transfer: an integrative analysis of research 1987–1992. Br.J.Obstet.Gynaecol., 102, 317–325.

Lazarus, R.S. (1975) The self-regulation of emotion. In Levi, L. (ed.), Emotions – Their Parameters and Measurement. Raven Press, New York, p. 47.

McNair, D. and Lorr, M. (1982) Profile of Mood States Bipolar Form (POMS-BI). Educational and Industrial Testing Service, San Diego.

Merari, D., Feldberg, D., Elizur, A. et al. (1992) Psychological and hormonal changes in the course of in vitro fertilization. J.Assist.Reprod.Genet., 9, 161–169.

Milad, M.P., Klock, S.C., Moses, S. et al. (1998) Stress and anxiety do not result in pregnancy wastage. Hum.Reprod., 13, 2296–2300.[Abstract]

Molloy, D., Doody, M.L. and Breen, T. (1995) Second time around: a study of patients seeking second assisted reproduction pregnancies. Fertil.Steril., 64, 546–551.[ISI][Medline]

Nelson, J.R., Corson, S.L., Batzer, F.R. et al. (1993) Predicting success of gamete intrafallopian transfer. Fertil.Steril., 60, 116–122.[ISI][Medline]

Reading, A.E. (1989) Decision making and in vitro fertilisation: the influence of emotional state. J.Psychosom.Obstet.Gynaecol., 10, 107–112.[ISI]

Shatford, L.A., Hearn, M.T., Yuzpe, A.A. et al. (1988) Psychological correlates of differential infertility diagnosis in an in vitro fertilization program. Am.J.Obstet.Gynecol., 158, 1099–1107.[ISI][Medline]

Shulman, A., Ben-Nun, I., Ghetler, Y. et al. (1993) Relationship between embryo morphology and implantation rate after in vitro fertilization treatment in conception cycles. Fertil.Steril., 60, 123–126.[ISI][Medline]

Slade, P., Emery, J. and Lieberman, B.A. (1997) A prospective, longitudinal study of emotions and relationships in in-vitro fertilization treatment. Hum.Reprod., 12, 183–190.[Abstract]

Spielberger, C.D., Gorsuch, R.L. and Lushene, R.E. (1970) STAI Manual for the State–Trait Anxiety Inventory (Self-evaluation Questionnaire). Consulting Psychologists Press Inc, Palo Alto, California.

Spielberger, C.D. and Rickman, R.L. (1990) Assessment of state and trait anxiety. In Sartorius, N., Andreoli, V., Cassano, G. et al. (eds), Anxiety: Psychobiological and Clinical Perspectives. Hemisphere Publishing Corporation, New York, p. 69.

Stoleru, S., Cornet, D., Vaugeois, P. et al. (1997) The influence of psychological factors on the outcome of the fertilization step of in vitro fertilization. J.Psychosom.Obstet.Gynecol., 18, 189–202.

Thiering, P., Beaurepaire, J., Jones, M. et al. (1993) Mood state as a predictor of treatment outcome after in vitro fertilization/embryo transfer technology (IVF/ET). J.Psychosom.Res., 37, 481–491.[ISI][Medline]

Wang, X.J., Warnes, G.M., Norman, R.J. et al. (1993) Gamete intra-fallopian transfer: outcome following the elective or non-elective replacement of two, three or four oocytes. Hum.Reprod., 8, 1231–1234.[Abstract]

Submitted on November 2, 1998; accepted on February 15, 1999.