- Email: [email protected]
Bulimia nervosa, childbirth, and psychopathology
Journal of Psychosomatic Research 55 (2003) 357 – 361
Bulimia nervosa, childbirth, and psychopathology Frances A. Cartera,*, Virginia V.W. McIntosha, Peter R. Joycea, Christopher M. Framptona, Cynthia M. Bulikb b
a Department of Psychological Medicine, Christchurch School of Medicine, Otago University, P.O. Box 4345, Christchurch, New Zealand Virginia Institute for Psychiatric and Behavioral Genetics, Department of Psychiatry, Virginia Commonwealth University, Richmond, VA, USA
Received 1 March 2002; accepted 26 November 2002
Abstract Objective: To examine whether having a baby following treatment for bulimia nervosa places women at increased risk for continuing or relapsing eating disorders or major depression. Methods: Subjects were women who had participated in a large randomized controlled trial evaluating cognitive behavior therapy for bulimia nervosa, who were prospectively followed-up over 5 years. At follow-up assessments (at least yearly), life charts were completed with patients and childbirth was recorded. The presence
of eating disorders and major depressive disorder was assessed using the Structured Interview for DSM-III-R. Results: Childbirth was not specifically associated with increased symptomatology. This was found for both eating disorders and major depression in the same year as childbirth and for the year following childbirth. Conclusion: Childbirth is not specifically associated with symptomatology following treatment for bulimia nervosa. D 2003 Elsevier Inc. All rights reserved.
Keywords: Bulimia nervosa; Childbirth; Postpartum; Eating disorders; Depression; Follow-up
Introduction Bulimia nervosa has been noted to commonly affect women at a peak age for reproductive functioning [1,2]. Bulimia nervosa and reproduction may interact in several ways. First, bulimia nervosa may adversely affect reproduction [1] (see Refs. [3– 5] for reviews). For example, women with bulimia nervosa have been reported to have elevated rates of menstrual and ovulatory abnormalities [6 –9], hyperemesis gravidarum [10,11], miscarriages and terminations of pregnancy [10], and investigations for infertility [10,12]. Second, reproduction may adversely affect bulimia nervosa. Retrospective studies examining patients with bulimia have reported that while pregnancy may be associated with a reduction in bulimic symptoms, a resurgence of symptoms is common in the postpartum period [2,13 – 15]. Few studies have examined the impact of childbirth on depression in a sample of women with bulimia nervosa [2,10]. This is surprising given the known high rates of comorbidity between bulimia nervosa and affective disorders [16 – 21]. Morgan et al. [2] found that overall, postpartum depression * Corresponding author. Tel.: +64-3-372-0400; fax: +64-3-372-0407. E-mail address: [email protected] (F.A. Carter). 0022-3999/03/$ – see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0022-3999(02)00641-4
was suggested in approximately one-third of their sample. Abraham [10] reported similar rates of treatment for postpartum depression. An active eating disorder [10] and a history of alcohol misuse and anorexia nervosa [2] were associated with increased risk of postpartum depression. It is unknown what the impact of treatment for bulimia nervosa may have on reproduction and subsequent psychopathology. Those studies that have reported treatment status have involved participants who were untreated (e.g., Ref. [13]), partially treated (e.g., Ref. [14]), or had been treated a long time ago (e.g., 10 – 15 years ago [10]). The present study sought to examine whether having a baby following treatment for bulimia nervosa places women at increased risk for continuing or relapsing eating disorders or major depression. This issue was examined within the context of a large randomized controlled trial evaluating cognitive behavior therapy for bulimia nervosa [22], with follow-up over 5 years.
Method Participants in the original treatment study were women aged between 17 and 45 years with a current, primary
358
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
nervosa, anorexia nervosa, eating disorders not otherwise specified, major depressive syndrome, substance-use disorders, and social phobia. Diagnoses were made for the past month and for the past year. The present study reports data for the past year. These diagnoses indicate that all diagnostic criteria were met at any stage over the past year, including in the past month. A summary variable ‘‘eating disorder (any)’’ was calculated (yes/no). The present study utilizes this summary variable in order to minimize repeated comparisons and avoid problems with small numbers with diagnoses of anorexia nervosa. At each follow-up assessment, data were also collected concerning patients’ menstrual and weight history over the previous year, and pregnancy or childbirth was noted. Life charts were also completed whereby patients identified key life events (e.g., becoming pregnant or having a baby), which were used as anchor points to assist them with recalling bulimic behaviors at different time points over the previous year.
Results Fig. 1 shows a summary of patient flow. Overall, attendance at follow-up assessments ranged from 66% to 84% of patients over the 5-year follow-up period. Over the 5 years of follow-up, 43 babies were born to 32 women (range: 0– 3 babies; year 1: 6, year 2: 11, year 3: 8, year 4: 11, and year 5: 7). The nine women who had more than one child over follow-up were compared with the women who had one child over follow-up in terms of the rate of eating disorder (any) and major depressive disorder diagnoses. No significant differences emerged between these groups using chi-square tests. Therefore, all subsequent analyses group women who had babies over follow-up together, regardless of how many babies they had over this time.
Fig. 1. Overview of study design and patient flow at assessment points.
diagnosis of bulimia nervosa [23]. Exclusion criteria were current anorexia nervosa, current obesity (body mass index [BMI] > 30 kg/m2), current severe major depression with severe suicidal ideation or requiring immediate treatment with antidepressants, current severe medical illness or severe medical complications of bulimia nervosa, or the current use of psychoactive medication and unwillingness to undergo a supervised drug wash-out period. The original sample have been described in detail elsewhere [22]. Patients were assessed at pretreatment, mid-treatment, end-treatment, 6 months posttreatment, and at yearly intervals posttreatment for 5 years. Fig. 1 shows an overview of the study design. Over the follow-up period, the presence of the following Axis I disorders was determined using the Structured Clinical Interview for DSM-III-R [24]: bulimia
Co-occurrences of having a baby and symptomatology The impact of having a baby and experiencing symptomatology in the same year was examined (cooccurrences). Table 1 shows co-occurrences of having had a baby and suffered an eating disorder (any) and/or major depression. With one exception (year 4 of follow-up for major depression), women who had had a baby had
Table 1 Co-occurrence of having had a baby and an eating disorder (any) and/or major depression in the same year, over 5-year follow-up Year of follow-up 1
2
3
4
5
Diagnoses
No baby
Baby
P
No baby
Baby
P
No baby
Baby
P
No baby
Baby
P
No baby
Baby
P
Eating disorder (any) Major depression
54% (52/96) 46% (44/96)
17% (1/6) 17% (1/6)
.10
52% (41/79) 30% (24/79)
40% (4/10) 18% (2/11)
.52
43% (45/105) 37% (39/105)
13% (1/8) 0% (0/8)
.14
41% (40/98) 36% (35/98)
30% (3/10) 46% (5/11)
.74
35% (36/102) 45% (46/102)
29% (2/7) 29% (2/7)
1.00
.23
.50
.05
.53
.46
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
359
Table 2 Lagged occurrence of having had a baby and then an eating disorder (any) and/or major depression in the following year, over 5-year follow-up Year of follow-up 2 Diagnoses Eating disorder (any) Major depression
3
4
No baby (in year 1)
Baby (in year 1)
P
No baby (in year 2)
Baby (in year 2)
54% (41/76)
0% (0/6)
.03
41% (30/74)
27% (3/11)
31% (24/77)
17% (1/6)
.66
31% (23/74)
27% (3/11)
lower rates of eating disorders (any) and major depression than women who had not had a baby, at every year over follow-up. These differences were not statistically significant using chi-square tests (year 3 of follow-up for major depression approaches significance). Because of the small number of women having babies in any single year, a summary variable was calculated in order to improve the power available to examine the impact of childbirth on symptomatology. This variable represents the proportion of baby-linked (versus nonlinked) co-occurrences of having a baby and having an eating disorder and/or major depression. Analyses using Wilcoxon signed ranks tests show no significant differences between baby-linked versus nonlinked co-occurrences of eating disorders (any) ( P = .83) or major depression ( P = .35). One of the difficulties with examining co-occurrences in the ‘‘same year’’ is that this does not provide information about the sequence of childbirth and symptomatology within that year. In addition, it is also possible that if childbirth does have an adverse impact on psychopathology, than this impact may be delayed. For these reasons, we also examined ‘‘lagged occurrences’’ of having a baby and experiencing symptomatology.
5
No baby (in year 3)
Baby (in year 3)
P
No baby (in year 4)
Baby (in year 4)
.52
42% (39/94)
13% (1/8)
.14
37% (34/92)
27% (3/11)
.74
1.00
37% (35/95)
13% (1/8)
.26
46% (42/92)
46% (5/11)
1.00
P
P
signed ranks tests show no significant differences between baby-linked versus nonlinked lagged occurrences of eating disorders (any) or major depression over follow-up. Symptomatology in pre-baby years versus post-baby years Finally, the impact of childbirth on symptomatology over 5-year follow-up was examined by comparing symptomatology before having a baby with symptomatology after having a baby. Specifically, the proportion of prebaby years when an eating disorder (any) and/or major depression were present was compared with the proportion of post-baby years when an eating disorder and/or major depression were present over follow-up. Because pre-baby data were available from year 2 of follow-up, and post-baby data were available up until year 4 of follow-up, pre- and post-baby data were only available for 22 participants. Analyses using Wilcoxon signed ranks tests show no significant differences between pre- and post-baby episodes of eating disorders (any) ( P =.55) or major depression ( P =.48).
Discussion Lagged occurrences of having a baby and symptomatology The impact of having a baby and experiencing symptomatology in the year following childbirth was examined (lagged occurrences). Table 2 shows lagged occurrences of having a baby and then suffering an eating disorder (any) and/ or major depression. At each point assessed over 5-year follow-up, women had either the same or lower rates of eating disorders (any) and major depression in the year following childbirth than women who had not had a baby in the previous year. With one exception (rates of eating disorder diagnoses [any] in year 2 of follow-up), these differences were not statistically significant using chi-square tests. Again, because of the small number of women having babies in any single year, a summary variable was calculated in order to improve the power available to examine the impact of childbirth on symptomatology. This variable represents the proportion of baby-linked (versus nonlinked) lagged occurrences of having a baby and having an eating disorder and/or major depression. Analyses using Wilcoxon
The present study prospectively followed women over 5 years following treatment for bulimia nervosa and evaluated the impact of childbirth on symptomatology. We found that childbirth was not specifically associated with elevated rates of symptomatology in this sample. In fact, a trend was found for participants who had had a child to be less symptomatic than those who had not had a child. This trend was found for both eating disorders and major depression, in the same year as childbirth and for the year following childbirth. An important difference between the current study and many previous studies is that participants had received active treatment for their bulimia nervosa before childbirth in the present study. While the present study found that for this sample childbirth was not specifically associated with increased symptomatology, the sample as a whole had elevated rates of symptomatology. Rates of eating disorders (any) (overall rate of 43% over previous year across 5 years) were consistent with other treatment outcome studies at long term
360
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
follow-up (see Ref. [25] for a review). Participants also had elevated rates (overall rate of 40% over previous year across 5 years) of major depression in the year of childbirth and the year following childbirth in comparison with studies involving heterogeneous samples [26 – 36]. This is consistent with Abraham [10] and Morgan et al. [2] who reported similar rates of postpartum depression in a sample of women previously treated for bulimia nervosa. A number of methodological factors warrant consideration. First, although bulimia nervosa has been noted to commonly affect women at a peak age for reproductive functioning [1,2], having a baby in any given year was still a relatively rare event for participants in the study. Thus, power was limited for addressing the impact of childbirth on symptomatology in any single year. However, a lack of power would make these analyses vulnerable to a Type II error (i.e., to conclude that there is no association between the variables when in fact there is an association) [37]. As stated previously, the trend in the present study was to find that childbirth was associated with less symptomatology. The use of ‘‘summary variables’’ improved the power available to examine the relation between childbirth and symptomatology. The present study is the largest known study to have examined the relation between childbirth and symptomatology among women who were prospectively recruited with bulimia nervosa. Second, the data provide relatively crude information regarding the timing of childbirth and the onset of symptomatology. For this reason, we examined the relation between childbirth and symptomatology in a variety of ways using data from the same year and the year following childbirth. All of these analyses show a consistent pattern of results. Third, at each assessment over 5-year follow-up, diagnoses were made for the past month and for the past year. Retrospective diagnoses are always vulnerable to error. However, it is unlikely that any ‘‘errors’’ were associated with childbirth/nonchildbirth in a systematic way. Therefore, it is unlikely that these would have biased the current results. Fourth, it is possible, although we believe not likely, that some women who presented for assessment did not tell us that they had had a baby. What explanations exist for the current results? It is possible that having a baby may have had a positive impact on symptomatology. Alternatively, it may be that those women who had less symptomatology following treatment were more likely to go on to have a baby. As stated previously, active bulimia nervosa may adversely affect reproduction [1] (see Refs. [3– 5] for reviews). For example, Abraham [10] notes that the return of menstruation requires women to cease bulimic behaviors (e.g., self-induced vomiting, intermittent starvation, and excessive exercise) and maintain a BMI 19. A recent analysis of our data confirms that functioning poorly following treatment for bulimia nervosa, in combination with demographic variables, is strongly predictive of not going on to have a baby over 5-year follow-up [38].
In conclusion, the results of the present study show that childbirth is not specifically associated with symptomatology following treatment for bulimia nervosa. The present results require replication, and potential explanations for these results warrant further evaluation.
References [1] Morgan JF. Eating disorders and reproduction. Aust N Z J Obstet Gynaecol 1999;39:167 – 73. [2] Morgan JF, Lacey JH, Sedgwick PM. Impact of pregnancy on bulimia nervosa. Br J Psychiatry 1999;174:135 – 40. [3] Franko DL, Walton BE. Pregnancy and eating disorders: a review and clinical implications. Int J Eat Disord 1993;13:41 – 7. [4] Stewart DE. Reproductive functions in eating disorders. Ann Med 1992;24:287 – 91. [5] Morrill ES, Nickols-Richardson HM. Bulimia nervosa during pregnancy: a review. J Am Diet Assoc 2001;101:448 – 54. [6] Copeland PM, Herzog DB. Menstrual abnormalities. In: Hudson JI, Pope HG, editors. The psychobiology of bulimia. Washington (DC): American Psychiatric Press, 1987. pp. 31 – 54. [7] Devlin MJ, Walsh BT, Katz JL, Roose SP, Linkie DM, Wright L, Vande Wiele R, Glassman AH. Hypothalamic – pituitary – gonadal function in anorexia nervosa and bulimia. Psychiatry Res 1989; 28:11 – 24. [8] Pirke KM, Dogs M, Fichter MM, Tuschl RJ. Gonadotrophins, oestradiol and progesterone during the menstrual cycle in bulimia nervosa. Clin Endocrinol (Oxf ) 1988;29:265 – 70. [9] Pirke KM, Schweiger U, Lemmel W, Krieg JC, Berger M. The influence of dieting on the menstrual cycle of healthy young women. J Clin Endocrinol Metab 1985;60:1174 – 9. [10] Abraham S. Sexuality and reproduction in bulimia nervosa patients over 10 years. J Psychosom Res 1998;44:491 – 502. [11] Stewart DE, MacDonald OL. Hyperemesis gravidarum and eating disorders in pregnancy. In: Abraham S, Llewellyn-Jones D, editors. Eating disorders and disordered eating. Sydney: Ashwood House, 1987. pp. 52 – 5. [12] Stewart DE, Robinson E, Goldbloom DS, Wright C. Infertility and eating disorders. Am J Obstet Gynecol 1990;163:1196 – 9. [13] Lacey JH, Smith G. Bulimia nervosa. The impact of pregnancy on mother and baby. Br J Psychiatry 1987;150:777 – 81. [14] Lemberg R, Phillips J. The impact of pregnancy on anorexia nervosa and bulimia. Int J Eat Disord 1989;8:285 – 95. [15] Willis DC, Rand CS. Pregnancy in bulimic women. Obstet Gynecol 1988;71:708 – 10. [16] Brewerton TD, Lydiard RB, Herzog DB, Brotman AW, O’Neil PM, Ballenger JC. Comorbidity of Axis I psychiatric disorders in bulimia nervosa. J Clin Psychiatry 1995;56(2):77 – 80. [17] Mitchell JE, Specker SM, de Zwann M. Comorbidity and medical complications of bulimia nervosa. J Clin Psychiatry 1991;52:13 – 20 (Supplement). [18] Wonderlich SA, Mitchell JE. Eating disorders and comorbidity: empirical, conceptual, and clinical implications. Psychopharmacol Bull 1997;33:381 – 90. [19] Laessle RG, Schweiger U, Fitcher MM, Pirke KM. Eating disorders and depression: psychobiological findings in bulimia and anorexia nervosa. In: Pirke KM, Vandereycken W, Ploog D, editors. The psychobiology of bulimia nervosa. Berlin: Springer-Verlag, 1988. pp. 90 – 100. [20] Powers PS, Coovert DL, Brightwell DR, Stevens BA. Other psychiatric disorders among bulimic patients. Compr Psychiatry 1988;29: 503 – 8. [21] Lilenfeld LR, Kaye WH, Greeno CG, Merikangas KR, Plotnicov K, Pollice C, Rao R, Strober M, Bulik CM, Nagy L. Psychiatric disorders
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
[22]
[23]
[24]
[25] [26] [27]
[28] [29]
in women with bulimia nervosa and their first-degree relatives: effects of comorbid substance dependence. Int J Eat Disord 1997;22:253 – 64. Bulik CM, Sullivan PF, Carter FA, McIntosh VV, Joyce PR. The role of exposure with response prevention in the cognitive behavioural therapy for bulimia nervosa. Psychol Med 1998;28:611 – 23. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 3rd ed, revised. Washington (DC): American Psychiatric Association, 1987. Spitzer R, Williams J. Revised diagnostic criteria and a new structured interview for diagnosing anxiety disorders. Journal of Psychiatric Research 1988;22(Suppl 1):55 – 85. Keel PK, Mitchell JE. Outcome in bulimia nervosa. Am J Psychiatry 1997;154:313 – 21. Campbell SB, Cohn JF. Prevalence and correlates of postpartum depression in first-time mothers. J Abnorm Psychol 1991;100:594 – 9. Cooper PJ, Campbell EA, Day A, Kennerley H, Bond A. Nonpsychotic psychiatric disorder after childbirth. A prospective study of prevalence, incidence, course and nature. Br J Psychiatry 1988; 152:799 – 806. Cox JL, Connor Y, Kendell RE. Prospective study of the psychiatric disorders of childbirth. Br J Psychiatry 1982;140:111 – 7. Da Costa D, Larouche J, Dritsa M, Brender W. Psychosocial correlates of prepartum and postpartum depressed mood. J Affect Disord 2000; 59:31 – 40.
361
[30] Evans J, Heron J, Francomb H, Oke S, Golding J. Cohort study of depressed mood during pregnancy and after childbirth. Br Med J 2001;323:257 – 60. [31] Gotlib IH, Whiffen VE, Mount JH, Milne K, Cordy NI. Prevalence rates and demographic characteristics associated with depression in pregnancy and the postpartum. J Consult Clin Psychol 1989;57: 269 – 74. [32] Kumar R, Robson KM. A prospective study of emotional disorders in childbearing women. Br J Psychiatry 1984;144:35 – 47. [33] McGill H, Burrows VL, Holland LA, Langer HJ, Sweet MA. Postnatal depression: a Christchurch study. N Z Med J 1995;108:162 – 5. [34] O’Hara MW. Social support, life events, and depression during pregnancy and the puerperium. Arch Gen Psychiatry 1986;43:569 – 73. [35] Robinson S, Young J. Screening for depression and anxiety in the post-natal period: acceptance or rejection of a subsequent treatment offer. Aust N Z J Psychiatry 1982;16:47 – 51. [36] Watson JP, Elliott SA, Rugg AJ, Brough DI. Psychiatric disorder in pregnancy and the first postnatal year. Br J Psychiatry 1984;144: 453 – 62. [37] Streiner D. Sample size and power in psychiatric research. Can J Psychiatry 1990;35:616 – 20. [38] Carter FA, McIntosh VVW, Frampton CM, Joyce PR, Bulik CM. Predictors of childbirth following treatment for bulimia nervosa. Int J Eat Disord [in press].
Bulimia nervosa, childbirth, and psychopathology Frances A. Cartera,*, Virginia V.W. McIntosha, Peter R. Joycea, Christopher M. Framptona, Cynthia M. Bulikb b
a Department of Psychological Medicine, Christchurch School of Medicine, Otago University, P.O. Box 4345, Christchurch, New Zealand Virginia Institute for Psychiatric and Behavioral Genetics, Department of Psychiatry, Virginia Commonwealth University, Richmond, VA, USA
Received 1 March 2002; accepted 26 November 2002
Abstract Objective: To examine whether having a baby following treatment for bulimia nervosa places women at increased risk for continuing or relapsing eating disorders or major depression. Methods: Subjects were women who had participated in a large randomized controlled trial evaluating cognitive behavior therapy for bulimia nervosa, who were prospectively followed-up over 5 years. At follow-up assessments (at least yearly), life charts were completed with patients and childbirth was recorded. The presence
of eating disorders and major depressive disorder was assessed using the Structured Interview for DSM-III-R. Results: Childbirth was not specifically associated with increased symptomatology. This was found for both eating disorders and major depression in the same year as childbirth and for the year following childbirth. Conclusion: Childbirth is not specifically associated with symptomatology following treatment for bulimia nervosa. D 2003 Elsevier Inc. All rights reserved.
Keywords: Bulimia nervosa; Childbirth; Postpartum; Eating disorders; Depression; Follow-up
Introduction Bulimia nervosa has been noted to commonly affect women at a peak age for reproductive functioning [1,2]. Bulimia nervosa and reproduction may interact in several ways. First, bulimia nervosa may adversely affect reproduction [1] (see Refs. [3– 5] for reviews). For example, women with bulimia nervosa have been reported to have elevated rates of menstrual and ovulatory abnormalities [6 –9], hyperemesis gravidarum [10,11], miscarriages and terminations of pregnancy [10], and investigations for infertility [10,12]. Second, reproduction may adversely affect bulimia nervosa. Retrospective studies examining patients with bulimia have reported that while pregnancy may be associated with a reduction in bulimic symptoms, a resurgence of symptoms is common in the postpartum period [2,13 – 15]. Few studies have examined the impact of childbirth on depression in a sample of women with bulimia nervosa [2,10]. This is surprising given the known high rates of comorbidity between bulimia nervosa and affective disorders [16 – 21]. Morgan et al. [2] found that overall, postpartum depression * Corresponding author. Tel.: +64-3-372-0400; fax: +64-3-372-0407. E-mail address: [email protected] (F.A. Carter). 0022-3999/03/$ – see front matter D 2003 Elsevier Inc. All rights reserved. doi:10.1016/S0022-3999(02)00641-4
was suggested in approximately one-third of their sample. Abraham [10] reported similar rates of treatment for postpartum depression. An active eating disorder [10] and a history of alcohol misuse and anorexia nervosa [2] were associated with increased risk of postpartum depression. It is unknown what the impact of treatment for bulimia nervosa may have on reproduction and subsequent psychopathology. Those studies that have reported treatment status have involved participants who were untreated (e.g., Ref. [13]), partially treated (e.g., Ref. [14]), or had been treated a long time ago (e.g., 10 – 15 years ago [10]). The present study sought to examine whether having a baby following treatment for bulimia nervosa places women at increased risk for continuing or relapsing eating disorders or major depression. This issue was examined within the context of a large randomized controlled trial evaluating cognitive behavior therapy for bulimia nervosa [22], with follow-up over 5 years.
Method Participants in the original treatment study were women aged between 17 and 45 years with a current, primary
358
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
nervosa, anorexia nervosa, eating disorders not otherwise specified, major depressive syndrome, substance-use disorders, and social phobia. Diagnoses were made for the past month and for the past year. The present study reports data for the past year. These diagnoses indicate that all diagnostic criteria were met at any stage over the past year, including in the past month. A summary variable ‘‘eating disorder (any)’’ was calculated (yes/no). The present study utilizes this summary variable in order to minimize repeated comparisons and avoid problems with small numbers with diagnoses of anorexia nervosa. At each follow-up assessment, data were also collected concerning patients’ menstrual and weight history over the previous year, and pregnancy or childbirth was noted. Life charts were also completed whereby patients identified key life events (e.g., becoming pregnant or having a baby), which were used as anchor points to assist them with recalling bulimic behaviors at different time points over the previous year.
Results Fig. 1 shows a summary of patient flow. Overall, attendance at follow-up assessments ranged from 66% to 84% of patients over the 5-year follow-up period. Over the 5 years of follow-up, 43 babies were born to 32 women (range: 0– 3 babies; year 1: 6, year 2: 11, year 3: 8, year 4: 11, and year 5: 7). The nine women who had more than one child over follow-up were compared with the women who had one child over follow-up in terms of the rate of eating disorder (any) and major depressive disorder diagnoses. No significant differences emerged between these groups using chi-square tests. Therefore, all subsequent analyses group women who had babies over follow-up together, regardless of how many babies they had over this time.
Fig. 1. Overview of study design and patient flow at assessment points.
diagnosis of bulimia nervosa [23]. Exclusion criteria were current anorexia nervosa, current obesity (body mass index [BMI] > 30 kg/m2), current severe major depression with severe suicidal ideation or requiring immediate treatment with antidepressants, current severe medical illness or severe medical complications of bulimia nervosa, or the current use of psychoactive medication and unwillingness to undergo a supervised drug wash-out period. The original sample have been described in detail elsewhere [22]. Patients were assessed at pretreatment, mid-treatment, end-treatment, 6 months posttreatment, and at yearly intervals posttreatment for 5 years. Fig. 1 shows an overview of the study design. Over the follow-up period, the presence of the following Axis I disorders was determined using the Structured Clinical Interview for DSM-III-R [24]: bulimia
Co-occurrences of having a baby and symptomatology The impact of having a baby and experiencing symptomatology in the same year was examined (cooccurrences). Table 1 shows co-occurrences of having had a baby and suffered an eating disorder (any) and/or major depression. With one exception (year 4 of follow-up for major depression), women who had had a baby had
Table 1 Co-occurrence of having had a baby and an eating disorder (any) and/or major depression in the same year, over 5-year follow-up Year of follow-up 1
2
3
4
5
Diagnoses
No baby
Baby
P
No baby
Baby
P
No baby
Baby
P
No baby
Baby
P
No baby
Baby
P
Eating disorder (any) Major depression
54% (52/96) 46% (44/96)
17% (1/6) 17% (1/6)
.10
52% (41/79) 30% (24/79)
40% (4/10) 18% (2/11)
.52
43% (45/105) 37% (39/105)
13% (1/8) 0% (0/8)
.14
41% (40/98) 36% (35/98)
30% (3/10) 46% (5/11)
.74
35% (36/102) 45% (46/102)
29% (2/7) 29% (2/7)
1.00
.23
.50
.05
.53
.46
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
359
Table 2 Lagged occurrence of having had a baby and then an eating disorder (any) and/or major depression in the following year, over 5-year follow-up Year of follow-up 2 Diagnoses Eating disorder (any) Major depression
3
4
No baby (in year 1)
Baby (in year 1)
P
No baby (in year 2)
Baby (in year 2)
54% (41/76)
0% (0/6)
.03
41% (30/74)
27% (3/11)
31% (24/77)
17% (1/6)
.66
31% (23/74)
27% (3/11)
lower rates of eating disorders (any) and major depression than women who had not had a baby, at every year over follow-up. These differences were not statistically significant using chi-square tests (year 3 of follow-up for major depression approaches significance). Because of the small number of women having babies in any single year, a summary variable was calculated in order to improve the power available to examine the impact of childbirth on symptomatology. This variable represents the proportion of baby-linked (versus nonlinked) co-occurrences of having a baby and having an eating disorder and/or major depression. Analyses using Wilcoxon signed ranks tests show no significant differences between baby-linked versus nonlinked co-occurrences of eating disorders (any) ( P = .83) or major depression ( P = .35). One of the difficulties with examining co-occurrences in the ‘‘same year’’ is that this does not provide information about the sequence of childbirth and symptomatology within that year. In addition, it is also possible that if childbirth does have an adverse impact on psychopathology, than this impact may be delayed. For these reasons, we also examined ‘‘lagged occurrences’’ of having a baby and experiencing symptomatology.
5
No baby (in year 3)
Baby (in year 3)
P
No baby (in year 4)
Baby (in year 4)
.52
42% (39/94)
13% (1/8)
.14
37% (34/92)
27% (3/11)
.74
1.00
37% (35/95)
13% (1/8)
.26
46% (42/92)
46% (5/11)
1.00
P
P
signed ranks tests show no significant differences between baby-linked versus nonlinked lagged occurrences of eating disorders (any) or major depression over follow-up. Symptomatology in pre-baby years versus post-baby years Finally, the impact of childbirth on symptomatology over 5-year follow-up was examined by comparing symptomatology before having a baby with symptomatology after having a baby. Specifically, the proportion of prebaby years when an eating disorder (any) and/or major depression were present was compared with the proportion of post-baby years when an eating disorder and/or major depression were present over follow-up. Because pre-baby data were available from year 2 of follow-up, and post-baby data were available up until year 4 of follow-up, pre- and post-baby data were only available for 22 participants. Analyses using Wilcoxon signed ranks tests show no significant differences between pre- and post-baby episodes of eating disorders (any) ( P =.55) or major depression ( P =.48).
Discussion Lagged occurrences of having a baby and symptomatology The impact of having a baby and experiencing symptomatology in the year following childbirth was examined (lagged occurrences). Table 2 shows lagged occurrences of having a baby and then suffering an eating disorder (any) and/ or major depression. At each point assessed over 5-year follow-up, women had either the same or lower rates of eating disorders (any) and major depression in the year following childbirth than women who had not had a baby in the previous year. With one exception (rates of eating disorder diagnoses [any] in year 2 of follow-up), these differences were not statistically significant using chi-square tests. Again, because of the small number of women having babies in any single year, a summary variable was calculated in order to improve the power available to examine the impact of childbirth on symptomatology. This variable represents the proportion of baby-linked (versus nonlinked) lagged occurrences of having a baby and having an eating disorder and/or major depression. Analyses using Wilcoxon
The present study prospectively followed women over 5 years following treatment for bulimia nervosa and evaluated the impact of childbirth on symptomatology. We found that childbirth was not specifically associated with elevated rates of symptomatology in this sample. In fact, a trend was found for participants who had had a child to be less symptomatic than those who had not had a child. This trend was found for both eating disorders and major depression, in the same year as childbirth and for the year following childbirth. An important difference between the current study and many previous studies is that participants had received active treatment for their bulimia nervosa before childbirth in the present study. While the present study found that for this sample childbirth was not specifically associated with increased symptomatology, the sample as a whole had elevated rates of symptomatology. Rates of eating disorders (any) (overall rate of 43% over previous year across 5 years) were consistent with other treatment outcome studies at long term
360
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
follow-up (see Ref. [25] for a review). Participants also had elevated rates (overall rate of 40% over previous year across 5 years) of major depression in the year of childbirth and the year following childbirth in comparison with studies involving heterogeneous samples [26 – 36]. This is consistent with Abraham [10] and Morgan et al. [2] who reported similar rates of postpartum depression in a sample of women previously treated for bulimia nervosa. A number of methodological factors warrant consideration. First, although bulimia nervosa has been noted to commonly affect women at a peak age for reproductive functioning [1,2], having a baby in any given year was still a relatively rare event for participants in the study. Thus, power was limited for addressing the impact of childbirth on symptomatology in any single year. However, a lack of power would make these analyses vulnerable to a Type II error (i.e., to conclude that there is no association between the variables when in fact there is an association) [37]. As stated previously, the trend in the present study was to find that childbirth was associated with less symptomatology. The use of ‘‘summary variables’’ improved the power available to examine the relation between childbirth and symptomatology. The present study is the largest known study to have examined the relation between childbirth and symptomatology among women who were prospectively recruited with bulimia nervosa. Second, the data provide relatively crude information regarding the timing of childbirth and the onset of symptomatology. For this reason, we examined the relation between childbirth and symptomatology in a variety of ways using data from the same year and the year following childbirth. All of these analyses show a consistent pattern of results. Third, at each assessment over 5-year follow-up, diagnoses were made for the past month and for the past year. Retrospective diagnoses are always vulnerable to error. However, it is unlikely that any ‘‘errors’’ were associated with childbirth/nonchildbirth in a systematic way. Therefore, it is unlikely that these would have biased the current results. Fourth, it is possible, although we believe not likely, that some women who presented for assessment did not tell us that they had had a baby. What explanations exist for the current results? It is possible that having a baby may have had a positive impact on symptomatology. Alternatively, it may be that those women who had less symptomatology following treatment were more likely to go on to have a baby. As stated previously, active bulimia nervosa may adversely affect reproduction [1] (see Refs. [3– 5] for reviews). For example, Abraham [10] notes that the return of menstruation requires women to cease bulimic behaviors (e.g., self-induced vomiting, intermittent starvation, and excessive exercise) and maintain a BMI 19. A recent analysis of our data confirms that functioning poorly following treatment for bulimia nervosa, in combination with demographic variables, is strongly predictive of not going on to have a baby over 5-year follow-up [38].
In conclusion, the results of the present study show that childbirth is not specifically associated with symptomatology following treatment for bulimia nervosa. The present results require replication, and potential explanations for these results warrant further evaluation.
References [1] Morgan JF. Eating disorders and reproduction. Aust N Z J Obstet Gynaecol 1999;39:167 – 73. [2] Morgan JF, Lacey JH, Sedgwick PM. Impact of pregnancy on bulimia nervosa. Br J Psychiatry 1999;174:135 – 40. [3] Franko DL, Walton BE. Pregnancy and eating disorders: a review and clinical implications. Int J Eat Disord 1993;13:41 – 7. [4] Stewart DE. Reproductive functions in eating disorders. Ann Med 1992;24:287 – 91. [5] Morrill ES, Nickols-Richardson HM. Bulimia nervosa during pregnancy: a review. J Am Diet Assoc 2001;101:448 – 54. [6] Copeland PM, Herzog DB. Menstrual abnormalities. In: Hudson JI, Pope HG, editors. The psychobiology of bulimia. Washington (DC): American Psychiatric Press, 1987. pp. 31 – 54. [7] Devlin MJ, Walsh BT, Katz JL, Roose SP, Linkie DM, Wright L, Vande Wiele R, Glassman AH. Hypothalamic – pituitary – gonadal function in anorexia nervosa and bulimia. Psychiatry Res 1989; 28:11 – 24. [8] Pirke KM, Dogs M, Fichter MM, Tuschl RJ. Gonadotrophins, oestradiol and progesterone during the menstrual cycle in bulimia nervosa. Clin Endocrinol (Oxf ) 1988;29:265 – 70. [9] Pirke KM, Schweiger U, Lemmel W, Krieg JC, Berger M. The influence of dieting on the menstrual cycle of healthy young women. J Clin Endocrinol Metab 1985;60:1174 – 9. [10] Abraham S. Sexuality and reproduction in bulimia nervosa patients over 10 years. J Psychosom Res 1998;44:491 – 502. [11] Stewart DE, MacDonald OL. Hyperemesis gravidarum and eating disorders in pregnancy. In: Abraham S, Llewellyn-Jones D, editors. Eating disorders and disordered eating. Sydney: Ashwood House, 1987. pp. 52 – 5. [12] Stewart DE, Robinson E, Goldbloom DS, Wright C. Infertility and eating disorders. Am J Obstet Gynecol 1990;163:1196 – 9. [13] Lacey JH, Smith G. Bulimia nervosa. The impact of pregnancy on mother and baby. Br J Psychiatry 1987;150:777 – 81. [14] Lemberg R, Phillips J. The impact of pregnancy on anorexia nervosa and bulimia. Int J Eat Disord 1989;8:285 – 95. [15] Willis DC, Rand CS. Pregnancy in bulimic women. Obstet Gynecol 1988;71:708 – 10. [16] Brewerton TD, Lydiard RB, Herzog DB, Brotman AW, O’Neil PM, Ballenger JC. Comorbidity of Axis I psychiatric disorders in bulimia nervosa. J Clin Psychiatry 1995;56(2):77 – 80. [17] Mitchell JE, Specker SM, de Zwann M. Comorbidity and medical complications of bulimia nervosa. J Clin Psychiatry 1991;52:13 – 20 (Supplement). [18] Wonderlich SA, Mitchell JE. Eating disorders and comorbidity: empirical, conceptual, and clinical implications. Psychopharmacol Bull 1997;33:381 – 90. [19] Laessle RG, Schweiger U, Fitcher MM, Pirke KM. Eating disorders and depression: psychobiological findings in bulimia and anorexia nervosa. In: Pirke KM, Vandereycken W, Ploog D, editors. The psychobiology of bulimia nervosa. Berlin: Springer-Verlag, 1988. pp. 90 – 100. [20] Powers PS, Coovert DL, Brightwell DR, Stevens BA. Other psychiatric disorders among bulimic patients. Compr Psychiatry 1988;29: 503 – 8. [21] Lilenfeld LR, Kaye WH, Greeno CG, Merikangas KR, Plotnicov K, Pollice C, Rao R, Strober M, Bulik CM, Nagy L. Psychiatric disorders
F.A. Carter et al. / Journal of Psychosomatic Research 55 (2003) 357–361
[22]
[23]
[24]
[25] [26] [27]
[28] [29]
in women with bulimia nervosa and their first-degree relatives: effects of comorbid substance dependence. Int J Eat Disord 1997;22:253 – 64. Bulik CM, Sullivan PF, Carter FA, McIntosh VV, Joyce PR. The role of exposure with response prevention in the cognitive behavioural therapy for bulimia nervosa. Psychol Med 1998;28:611 – 23. American Psychiatric Association. Diagnostic and statistical manual of mental disorders. 3rd ed, revised. Washington (DC): American Psychiatric Association, 1987. Spitzer R, Williams J. Revised diagnostic criteria and a new structured interview for diagnosing anxiety disorders. Journal of Psychiatric Research 1988;22(Suppl 1):55 – 85. Keel PK, Mitchell JE. Outcome in bulimia nervosa. Am J Psychiatry 1997;154:313 – 21. Campbell SB, Cohn JF. Prevalence and correlates of postpartum depression in first-time mothers. J Abnorm Psychol 1991;100:594 – 9. Cooper PJ, Campbell EA, Day A, Kennerley H, Bond A. Nonpsychotic psychiatric disorder after childbirth. A prospective study of prevalence, incidence, course and nature. Br J Psychiatry 1988; 152:799 – 806. Cox JL, Connor Y, Kendell RE. Prospective study of the psychiatric disorders of childbirth. Br J Psychiatry 1982;140:111 – 7. Da Costa D, Larouche J, Dritsa M, Brender W. Psychosocial correlates of prepartum and postpartum depressed mood. J Affect Disord 2000; 59:31 – 40.
361
[30] Evans J, Heron J, Francomb H, Oke S, Golding J. Cohort study of depressed mood during pregnancy and after childbirth. Br Med J 2001;323:257 – 60. [31] Gotlib IH, Whiffen VE, Mount JH, Milne K, Cordy NI. Prevalence rates and demographic characteristics associated with depression in pregnancy and the postpartum. J Consult Clin Psychol 1989;57: 269 – 74. [32] Kumar R, Robson KM. A prospective study of emotional disorders in childbearing women. Br J Psychiatry 1984;144:35 – 47. [33] McGill H, Burrows VL, Holland LA, Langer HJ, Sweet MA. Postnatal depression: a Christchurch study. N Z Med J 1995;108:162 – 5. [34] O’Hara MW. Social support, life events, and depression during pregnancy and the puerperium. Arch Gen Psychiatry 1986;43:569 – 73. [35] Robinson S, Young J. Screening for depression and anxiety in the post-natal period: acceptance or rejection of a subsequent treatment offer. Aust N Z J Psychiatry 1982;16:47 – 51. [36] Watson JP, Elliott SA, Rugg AJ, Brough DI. Psychiatric disorder in pregnancy and the first postnatal year. Br J Psychiatry 1984;144: 453 – 62. [37] Streiner D. Sample size and power in psychiatric research. Can J Psychiatry 1990;35:616 – 20. [38] Carter FA, McIntosh VVW, Frampton CM, Joyce PR, Bulik CM. Predictors of childbirth following treatment for bulimia nervosa. Int J Eat Disord [in press].