Depression, anxiety, and post-traumatic stress disorder symptoms after hyperemesis gravidarum: a prospective cohort study

Abstract

Objective

To determine the prevalence of depression, anxiety, and posttraumatic stress disorder (PTSD) years after hyperemesis gravidarum (HG) and its association with HG severity.

Material and methods

This prospective cohort study consisted of a follow-up of 215 women admitted for HG, who were eligible to participate in a randomized controlled trial and either declined or agreed to be randomized between 2013 and 2016 in 19 hospitals in the Netherlands. Participants completed the Hospital Anxiety and Depression Scale (HADS) six weeks postpartum and during follow-up and the PTSD checklist for DSM-5 (PCL-5) during follow-up. An anxiety or depression score ≥8 is indicative of an anxiety or depression disorder and a PCL-5 ≥ 31 indicative of PTSD. Measures of HG severity were symptom severity (PUQE-24: Pregnancy Unique Quantification of Emesis), weight change, duration of admissions, readmissions, and admissions after the first trimester.

Results

About 54/215 participants completed the HADS six weeks postpartum and 73/215 participants completed the follow-up questionnaire, on average 4.5 years later. Six weeks postpartum, 13 participants (24.1%) had an anxiety score ≥8 and 11 participants (20.4%) a depression score ≥8. During follow-up, 29 participants (39.7%) had an anxiety score ≥8, 20 participants (27.4%) a depression score ≥8, and 16 participants (21.9%) a PCL-5 ≥ 31.

Multivariable logistic regression analysis showed that for every additional point of the mean PUQE-24 three weeks after inclusion, the likelihood of having an anxiety score ≥8 and PCL-5 ≥ 31 at follow-up increased with OR 1.41 (95% CI: 1.10;1.79) and OR 1.49 (95% CI: 1.06;2.10) respectively.

Conclusion

Depression, anxiety, and PTSD symptoms are common years after HG occurred.

Keywords:

• Hyperemesis gravidarum
• depression disorder
• postpartum depression
• anxiety disorder
• posttraumatic stress disorder

Introduction

Hyperemesis gravidarum (HG) is a severe form of nausea and vomiting in pregnancy. Dehydration, electrolyte disturbances, or weight loss can necessitate hospital admission for intravenous rehydration or tube feeding [1]. A systematic review published in 2017 showed a higher incidence of depression and anxiety symptoms during pregnancy in women suffering from HG [2]. Some studies have suggested that psychiatric diagnoses predispose to HG [3], whereas others have argued that HG causes depression, anxiety as well as posttraumatic stress disorder (PTSD) symptoms [4–6]. The fact that HG symptom improvement has been associated with a reduction in anxiety and depression symptoms, supports the latter of the two hypotheses [7–9].

Increases in depression and PTSD symptoms after pregnancies complicated by HG have been reported [4–6,10,11]. There have been suggestions of a possible dose–response effect with higher depression and PTSD scores postpartum among women with increased HG symptoms or with a prolonged disease course, hinting at a causal relationship between HG and psychopathology that persists postpartum [4,5,11].

Altogether, there is a limited body of evidence on the size and strength of the association between HG symptom severity and depression, anxiety, and PTSD symptoms. Furthermore, recently the long term maternal mental health consequences of HG were identified by patients and clinicians as one of the top 10 HG priority research questions [12]. Therefore, in this study, we aim to prospectively determine the association between HG symptom severity and depression, anxiety, and PTSD symptoms in women years after HG diagnosis.

Material and methods

Study design MOTHER

Our study is a prospective follow-up study of the MOTHER (Maternal and Offspring outcomes after Treatment of HyperEmesis by Refeeding) study [13]. The MOTHER study included women admitted for HG between 5 and 20 weeks’ gestation and consisted of a randomized controlled trial (RCT) and associated observational cohort of women who were eligible for participation in the trial but who declined randomization. The RCT assessed whether early enteral tube feeding in addition to standard care for women admitted with HG improved maternal or perinatal outcomes. Since this was not the case, we were able to combine the RCT and cohort into one study population for this follow-up study. Between 2013 and 2016, the RCT included 115 women and the cohort another 100 women. The MOTHER study was approved by the research ethics committee of the Amsterdam UMC and registered at the Dutch trial register (NTR4197). More detailed information about the MOTHER study can be found in the previous published study protocol and earlier work [14,15].

Measures of HG severity of the index pregnancy

The MOTHER study prospectively collected detailed information about pregnancy and delivery, extracted by trained research staff from medical files. Data regarding medical history, including a history or ongoing disease of depressive, anxiety and/or eating disorders, were also collected from medical files. Participants self-reported their pre-pregnancy weight, ethnicity and highest completed level of education. If self-reported data were missing, these features were extracted from medical file where available.

In this study, the pregnancy in which women had participated in the MOTHER study was designated as the index pregnancy and measures of HG severity in this pregnancy were used as predictor variables in regression analysis. HG severity in the index pregnancy was assessed by symptom severity at inclusion, as measured by the self-reported 24-h Pregnancy Unique Quantification of Emesis (PUQE-24) score, and the mean PUQE-24 three weeks after inclusion, weight change, total duration of hospital admissions, being readmitted and being admitted after the first trimester. The PUQE-24 score could vary from 3 to 15 with a higher PUQE-24 score indicating more severe symptoms [16]. The mean PUQE-24 score in the first three weeks after inclusion was calculated by summing up available weekly PUQE-24 scores and dividing by the total number of weekly PUQE-24 scores available. Weight change was calculated by comparing weight at inclusion to pre-pregnancy weight.

Follow-up study procedures

MOTHER study participants who gave consent to be approached for follow-up studies were invited for this follow-up study by email. Participants completed a single online questionnaire. Reminder emails were sent on three occasions to complete the questionnaire. Informed consent was obtained separately for the MOTHER study and the present follow-up study. Ethical approval for the follow-up study was not necessary, according to the Medical Research Involving Human Subjects Act (W20_066 #20.094).

The questionnaire consisted of questions about maternal mental health, as well as subsequent pregnancies after inclusion in the MOTHER. We asked whether participants had suffered from HG again after the MOTHER study, defining HG as vomiting symptoms which occurred with either: multiple medication use, weight loss, hospital admission for HG, tube feeding or in which nausea and vomiting symptoms affected their life and/or work. The full follow-up questionnaire is enclosed as Appendix S1. More detailed findings on HG recurrence rates in subsequent pregnancies of this follow-up study are previously published [17].

Outcome measures: depression, anxiety, and PTSD symptoms

Depression and anxiety symptoms were self-reported at inclusion of the MOTHER study and six weeks postpartum by use of the Hospital Anxiety and Depression Scale (HADS) [18]. A higher HADS indicates more severe depression or anxiety symptoms: a depression or anxiety score greater than or equal to 8 is considered borderline and greater than or equal to 11 as abnormal [18]. Depression and anxiety symptoms were again assessed during the follow-up study using the HADS. PTSD symptoms were assessed during follow-up using the PTSD CheckList for DSM-5 (PCL-5) [19,20]. A higher PCL-5 score indicates more severe PTSD symptoms: a PCL-5 score greater or equal to 31 indicates PTSD. The follow-up questionnaire did not include questions about possible treatments for depression, anxiety, or PTSD symptoms. The questionnaire was available in both Dutch and English. The HADS has been validated in both languages [21] while the PCL-5 has only been validated in English [22].

Statistical analyses

To assess selective attrition, baseline characteristics and measures of HG severity and course of the index pregnancy were compared between women participating in the follow-up study and those who were lost to follow-up or declined participation in this study, using independent students t test, Mann–Whitney U test, and Chi-square test.

Univariable logistic and linear regression analysis was used to assess possible risk factors for depression, anxiety, and PTSD symptoms, including having an HG pregnancy between MOTHER and follow-up study participation. Subsequently, we performed univariable and multivariable logistic and linear regression analysis to assess the association between each of the measures of HG severity of the index pregnancy and dichotomous outcomes (HADS anxiety and depression score ≥8 and PCL-5 score ≥31) and continuous outcomes (total HADS and PCL-5 score). In multivariable logistic regression analysis, due to the small number of events, we were only able to adjust for a single confounder [23]. We opted to correct for the risk factor with the lowest p-value in univariable logistic regression analysis. In multivariable linear regression analysis, we were able to adjust for more confounders and included risk factors which were significantly associated in univariable linear regression analysis based on a p-value <.10 as confounders. Outcomes variables that were not normally distributed were log transformed, back transformed, and expressed in percentages of differences.

Lastly, we assessed differences in depression, anxiety, and PTSD symptoms between the treatment arms of the RCT. We performed an intention-to-treat, per protocol (receiving tube feeding within 3 d after randomization and continued for at least 7 d) and as treated analysis (receiving tube feeding within 7 d after randomization and continued for at least 7 d). Two-sided p-values <.05 were considered statistically significant. We used SPSS Statistics 26.0 for Windows (IBM Corp., Armonk, NY, USA) for all analyses.

Patient and public involvement

The Dutch HG patient support group Zwangerschapsmisselijkheid en Hyperemesis Gravidarum was involved in setting up the MOTHER and the follow-up study. One author (CD) is a patient representative and gave perspective on the interpretation of the results.

Results

Participants

From the 215 participating women in the MOTHER study, 54 women (25.1%) completed the HADS at six weeks postpartum between 2014 and 2016. (Supplement Figure 1) We approached 190 out of 215 women who gave consent to be contacted for follow-up studies and from whom we had an e-mail address available. Seventy-three women (34.0%) completed the follow-up questionnaire between March and May 2020, on average 4.5 years later. Baseline characteristics are presented in Table 1.

Table 1. Baseline characteristics and outcome measures of women included in this follow-up study.

Index pregnancy (during MOTHER study)
Baseline characteristics	N = 73
Age (years)	29.25 ± 4.62
Ethnicity
Western	52 (71.2%)
Non-western	12 (16.4%)
Education level
Primary or secondary	27 (37.0%)
Higher	29 (39.7%)
Primigravida at the time	26 (35.6%)
History of mental health disease^a	12 (16.4%)
HG in previous pregnancy^b	23 (48.9%)
HG in previous pregnancy requiring hospital admission^b	12 (25.5%)
Maternal outcomes
Weight change (kg)^c	−3.40 ± 3.85
PUQE-24 at inclusion	11.00 (9.00–13.00)
Mean PUQE-24 in the first 3 weeks after admission	9.19 ± 2.50
Total duration of hospital admissions (days)	5.00 (4.00–8.00)
Readmitted	24 (32.9%)
Admission after the first trimester	14 (19.2%)
HADS at inclusion	21.28 ± 6.65
Anxiety score ≥8^d	31 (50.8%)
Depression score ≥8^d	56 (91.8%)
Follow-up:
Depression, anxiety and PTSD symptoms
HADS 6 weeks postpartum^e	9.57 ± 6.24
Anxiety score ≥8^e	13 (24.1%)
Depression score ≥8^e	11 (20.4%)
HADS at follow-up study	10.00 (6.00–15.00)
Anxiety score ≥8	29 (39.7%)
Depression score ≥8	20 (27.4%)
PCL-5 score at follow-up study	13.00 (7.00–28.50)
PCL-5 score ≥31	16 (21.9%)
History of any traumatic event	23 (31.5%)
History of an obstetric traumatic event	3 (4.1%)
Women who experienced a subsequent pregnancy after MOTHER participation	35 (47.9%)
Suffered from HG again between participation of the MOTHER and follow-up study^f	30 (41.1%)

Data represented with mean ± SD, median (IQR), or frequency (%). ^aHistory of mental health disease can consist of a depressive, anxiety, PTSD, or eating disorder. ^bPercentage shown is frequency divided by multigravidas. ^cWeight change is weight at baseline minus prepregnancy weight and can be < 0 if women lost weight and can be > 0 if women gained weight. ^dPercentage shown is frequency divided by number of HADS at inclusion available (n = 61). ^eDifferent study group including 54 women. ^fHG defined as: if vomiting symptoms occurred with either multiple medication use, weight loss, hospital admission for HG, requiring tube feeding or whether nausea and vomiting symptoms affected their life and/or work.

HG: hyperemesis gravidarum, HADS: hospital anxiety and depression scale: a higher HADS indicates more severe anxiety or depression symptoms, PCL-5: PTSD checklist for the DSM 5, PTSD: posttraumatic stress disorder, PUQE-24: 24-h Pregnancy Unique Quantification of Emesis and nausea score: a higher PUQE-24 indicates more severe symptoms and can vary from 3 to 15.

As shown in Supplement Table S1, participants of the follow-up study were significantly more often of western ethnicity (71.2% vs. 50.0%, p = .01), had higher educational attainment (39.7% vs. 19.7%, p = .02), and had had higher PUQE-24 scores (11 (9–13) vs. 9 (7–12), p = .01) than women who were lost to follow-up or declined participation in the present study.

Depression and anxiety symptoms

At inclusion of the MOTHER study, while suffering from HG, 31 out of 61 participants (50.8%) had an anxiety score ≥8 and 56 out of 61 participants (91.8%) had a depression score ≥8 (Table 1). Six weeks postpartum, 13 out of 54 participants (24.1%) had an anxiety score ≥8 and 11 participants (20.4%) had a depression score ≥8. At follow-up, 29 out of 73 participants (39.7%) had an anxiety score ≥8 and 20 women (27.4%) had a depression score ≥8.

We were not able to identify risk factors that were associated with a depression or anxiety score ≥8 at six weeks postpartum in univariable logistic regression (Supplement Table S2). A history of any traumatic event (OR 3.63, 95% CI: 1.29;10.20) and having an HG pregnancy between MOTHER and follow-up participation (OR 2.74, 95% CI: 1.04;7.20) were associated with having an anxiety score ≥8 at follow-up (Supplement Table S2). Younger maternal age (OR 0.84, 95% CI: 0.73;0.96), a higher HADS at inclusion of the MOTHER study (OR 1.14, 95% CI: 1.02;1.27), and a history of a traumatic event (OR 3.08, 95% CI: 1.05;9.03) were associated with a depression score ≥8 at follow-up.

HG severity of the index pregnancy and depression and anxiety symptoms

Multivariable logistic regression analysis showed that for every additional point of the mean PUQE-24 score in the first three weeks after inclusion of the index pregnancy, the likelihood of having an anxiety score ≥8 at time of the follow-up study increased with OR 1.41 (95% CI: 1.10;1.79) (Table 2). None of the HG severity measures of the index pregnancy were associated with depression and anxiety scores six weeks postpartum or depression score ≥8 and total HADS at follow-up in both logistic and linear regression analysis (Table 2 and Supplement Table S3).

Table 2. Logistic regression analysis to assess the association between measures of HG severity in the index pregnancy and HADS at 6 weeks postpartum and HADS and PCL-5 score at follow-up study.

6 weeks postpartum	HADS 6 weeks postpartum: Anxiety score ≥8
	OR	95% CI
Weight change (kg)	0.95	0.82;1.10
PUQE-24 at inclusion	0.83	0.63;1.08
Mean PUQE-24 in the first 3 weeks after admission	0.87	0.68;1.12
Total duration of hospital admission(s) (days)	1.08	0.94;1.24
Readmitted	2.02	0.57;7.14
Admission after the first trimester	2.22	0.58;8.49
	HADS 6 weeks postpartum: Depression score ≥8
	OR	95% CI
Weight change (kg)	0.91	0.78;1.06
PUQE-24 at inclusion	0.76	0.57;1.02
Mean PUQE-24 in the first 3 weeks after admission	0.75	0.55;1.01
Total duration of hospital admission(s) (days)	1.12	0.97;1.29
Readmitted	1.28	0.34;4.84
Admission after the first trimester	1.89	0.46;7.78
Follow-up study	HADS at follow-up: Anxiety score ≥8
	Model 1		Model 2*
	OR	95% CI	OR	95% CI
Weight change (kg)	0.92	0.81;1.04	0.90	0.78;1.03
PUQE-24 at inclusion	1.03	0.86;1.23	1.07	0.88;1.30
Mean PUQE-24 in the first 3 weeks after admission	1.32	1.06;1.64	1.41	1.10;1.79
Total duration of hospital admission(s) (days)	0.95	0.86;1.05	0.95	0.86;1.06
Readmitted	0.58	0.22;1.52	0.63	0.23;1.73
Admission after the first trimester	0.81	0.24;2.72	0.74	0.21;2.63
	HADS at follow-up: Depression score ≥8
	Model 1		Model 2**
	OR	95% CI	OR	95% CI
Weight change (kg)	0.89	0.77;1.03	0.92	0.80;1.07
PUQE-24 at inclusion	1.02	0.84;1.24	1.09	0.87;1.38
Mean PUQE-24 in the first 3 weeks after admission	1.12	0.90;1.38	1.15	0.89;1.48
Total duration of hospital admission(s) (days)	0.98	0.91;1.07	0.96	0.86;1.08
Readmitted	0.65	0.22;1.89	0.30	0.08;1.11
Admission after the first trimester	1.63	0.47;5.63	1.34	0.32;5.68
	PCL-5 score at follow-up ≥31
	Model 1		Model 2**
	OR	95% CI	OR	95% CI
Weight change (kg)	0.95	0.83;1.10	0.99	0.82;1.18
PUQE-24 at inclusion	0.98	0.79;1.21	1.03	0.81;1.32
Mean PUQE-24 in the first 3 weeks after admission	1.30	1.01;1.67	1.49	1.06;2.10
Total duration of hospital admission(s) (days)	0.92	0.78;1.08	0.80	0.61;1.05
Readmitted	0.54	0.17;1.76	0.24	0.05;1.12
Admission after the first trimester	0.97	0.23;3.98	0.68	0.11;4.13

p-values < .05 are considered significant and marked in bold. Weight change during index pregnancy is weight at baseline minus prepregnancy weight and can be < 0 if women lost weight and > 0 if women gained weight. Logistic regression analysis: we were only able to perform multivariable logistic regression analysis for HADS and PCL-5 score at follow-up. Model 1: univariable regression analysis. Model 2: multivariable regression analysis: *Adjusted for history of any traumatic event, **Adjusted for HADS at inclusion during MOTHER study. HADS: hospital anxiety and depression scale. A higher HADS indicates more severe depression or anxiety symptoms whereas a depression or anxiety score ≥8 is considered borderline/abnormal. HG: hyperemesis gravidarum, PCL-5: post-traumatic stress disease (PTSD) checklist for the DSM-5. A higher PCL-5 score indicates more severe PTSD symptoms, whereas a PCL-5 score ≥31 indicates PTSD, PUQE: 24-h Pregnancy Unique Quantification of Emesis and nausea. A higher PUQE-24 score indicates more severe symptoms.

Depression and anxiety symptoms according to RCT treatment allocation

Among RCT participants, we did not find any differences in depression and anxiety symptoms at six weeks postpartum and at follow-up between women in the enteral tube feeding group and women in the standard care group in intention-to-treat, as treated and per protocol analysis (Supplement Tables S4–6).

PTSD symptoms

At follow-up, 16 out of 73 participants (21.9%) had a PCL-5 score ≥31, indicating PTSD (Table 1). A higher HADS at inclusion of the MOTHER study was associated with a PCL-5 score ≥31 in univariable logistic regression (OR 1.20, 95% CI: 1.04;1.29) and with higher PCL-5 scores in univariable linear regression (β 5.65, 95% CI: 1.31;10.08) (Supplement Table S2).

HG severity of the index pregnancy and PTSD symptoms

Multivariable logistic regression analysis showed that for every additional point of the mean PUQE-24 score in the first three weeks after inclusion in the index pregnancy, the likelihood of having a PCL-5 score ≥31 increased with OR 1.49 (95% CI: 1.06;2.10) (Table 2). Higher PUQE-24 scores in the first three weeks after inclusion in the index pregnancy were also associated with a higher PCL-5 score in multivariable linear regression analysis as shown in Supplement Table S3 (β 16.65, 95% CI: 5.34;29.05). None of the other measures of HG severity of the index pregnancy were associated with the PCL-5 score during follow-up.

PTSD symptoms according to RCT treatment allocation

We found higher PCL-5 scores at follow-up among those receiving enteral tube feeding compared with those receiving standard care (27.0 (18.0–43.0) vs. 12.0 (3.5–27.5), p = .046) in the as treated analysis (Supplement Table S4-6). There was no difference in the likelihood that women in the early enteral tube feeding group had PCL-5 scores ≥31 compared with women in the standard care group.

Discussion

Main findings

We found that, on average 4.5 years after having been admitted for HG, women were commonly affected by depression, anxiety, and PTSD symptoms. Depression (20%) and anxiety rates (24%) at six weeks postpartum and at follow-up (resp. 27% and 40%) were considerably higher than postpartum depression and anxiety rates reported in the general population (resp. 0.8–2.6% and 17%) [24,25]. Furthermore, we found that 22% of the women included in our study had PCL-5 scores indicative of probable PTSD, in line with earlier reports [6], and again, considerably higher than PTSD rates reported in the general postpartum population (0.3–4.0%) [24,26]. Importantly, our study suggests that higher vomiting scores in the index pregnancy were associated with an increased chance of meeting the diagnostic criteria for anxiety disorder and PTSD at follow-up.

Strengths and limitations

Prospectively, collected detailed information of the index pregnancy was one of the strengths of this study. The fact that we collected information on whether HG pregnancies had occurred between MOTHER and follow-up study participation was another strength. We used validated questionnaires to assess symptom severity during index pregnancy and to evaluate depression, anxiety, and PTSD symptoms. In addition, depression and anxiety symptoms were measured at three different moments.

The main limitation of this study was that it is a small sample study and therefore could have lacked sufficient power to detect differences. Only 34% (73/215) of the MOTHER participants completed the follow-up questionnaire, which made our study prone to selection bias and may hamper generalizability of our findings. Sensitivity analysis revealed that participants of the follow-up study had higher PUQE-24 scores in the index pregnancy than women who were lost to follow-up. Selective participation of more severely affected participants could partially explain the high depression, anxiety or PTSD rates in our study, but cannot explain the dose–effect association between increased PUQE-24 scores and increased mental health symptoms. The time-interval between participation of the MOTHER and follow-up study, which could have taken up to 6 years, could also have affected our results. Depression, anxiety, and PTSD symptoms may have improved over time, for example by receiving treatment, and therefore have led to an underestimation of symptoms [27,28]. Since we performed multiple statistical comparisons, it could be that some of our findings were due to chance. Unfortunately, no information on possible treatments for mental health disorders were collected in this follow-up study. Finally, we were not able to compare depression, anxiety, and PTSD rates between women with and without HG.

Interpretation

Depression rates in our study, at both six weeks postpartum and on average 4.5 years later, lie between the previously reported depression rates of 12% at 6 and 12 months postpartum by Kjeldgaard et al. [5] and 29% at 6 weeks postpartum by Mitchell-Jones et al. [29] These differences might be explained by the fact that Kjeldgaard et al. [5] included HG patients based on ICD codes instead of a clinical HG diagnosis, despite the fact that ICD codes have been demonstrated not to be reliably identify HG patients [30]. Misclassification of the disease could have led to underreporting of HG patients and in line depression rates. Conversely, due to self-selection participation in Mitchell-Jones et al. [29] and due to selective participation in our study, it could be that these study populations consisted of a more severe group of HG patients, which may have led to higher reported depression rates.

Two studies have reported that prolonged HG, persisting in or beyond the second trimester, increases the chance of depression symptoms postpartum, which is at odds with our study, since we did not find associations between measures of HG severity of the index pregnancy and depression symptoms [4,5]. Since these two studies had larger study populations (respectively 4.308 and 92.947 women), it could be that our study was simply too small to detect any significant associations.

Twenty two percent of the included women in our study had PCL-5 scores ≥31, indicative of probable PTSD, which is similar to the previously reported 18% of Christodoulou-Smith et al. [6] Their study did not specify which PTSD questionnaire was used and participants were retrieved through advertisement on a HG patient support group website with establishing HG pregnancies through self-reports, which could have led to an overestimation of PTSD rates. Both studies however provide evidence that PTSD symptoms are common in women who suffered from HG. There are several effective treatments available for PTSD, also during pregnancy [31].

A more recent published study from Kjeldgaard et al. [11] evaluated PTSD symptoms in the Norwegian Mother and Child Cohort by the use of the Impact of Event Score, and reported that women with HG had higher PTSD scores at eight weeks postpartum compared with women with no, mild, or severe nausea; an association that remained after adjusting for having a depression or anxiety disorder in their medical history. These findings are similar to ours. We found that higher vomiting scores were associated with higher PTSD and anxiety symptoms at follow-up. Unfortunately, our study design hampers our ability to draw any firm conclusions about the causal direction of this association. However, it is important to highlight the fact that our study, together with previous published studies, found a dose–effect response between HG symptom severity and depression, anxiety, and PTSD symptoms. Together with the fact that only 16% of included patients in this study had a medical history of a mental health disease, these findings support the notion that HG itself leads to depression, anxiety, or PTSD symptoms, instead of women with a preexistent psychiatric illness being more predisposed to develop HG. This is consistent with qualitative studies stating that the burden of HG leads to developing psychological symptoms instead of being the cause of the disease [32]. This notion is further supported by the fact that in our study an additional pregnancy affected by HG between the index pregnancy and follow-up participation more than doubled the odds of having an anxiety disorder at follow-up, a further suggestion of a dose–response effect with increased HG burden negatively impacting future mental health.

Conclusion

Taken together, our study confirms that anxiety, depression, and PTSD symptoms are common in women previously admitted for HG. Moreover, our study suggests that an increased burden of HG, either as evident from higher symptom scores or higher total number of HG affected pregnancies, are at an increased risk of developing an anxiety disorder or PTSD. Future studies should confirm whether better treatment of HG can prevent or improve depression, anxiety and PTSD symptoms.

Ethics approval

The MOTHER trial was approved by the Research Ethics Committee of the Amsterdam University Medical Centers (UMC), location AMC, on 3 April 2013. The MOTHER study was also registered at www.trialregister.nl by NTR4197. According to the Medical Research Involving Human Subjects Act, ethical approval for the follow-up study was not necessary (reference number W20_066 #20.094).

Author contributions

IJG, TJR, and RCP conceived and conducted the MOTHER study. KN and RCP conceived and conducted the follow-up study. LMvdM helped develop the online survey tool. JMJB, CR-S, HAB, DPvdH, WMH, AH, GK, SK, JOEHvL, JL, FvdM, DP, M-JP, PJP, LvRF, RJR, HCJS, TV, BWM, MHK, IJG, and RCP recruited participants for the original MOTHER study and collected data. KN performed the statistical analyses, under supervision of RCP and RvE, and drafted the manuscript. CD is a patient representative and gave perspective on the interpretation of the results. All authors (LvdM, CD, JMJB, CR-S, RvE, HAB, DPvdH, WMH, AH, GK, SK, JOEHvL, JL, FvdM, DP, M-JP, PJP, LvRF, RJR, HCJS, TV, BWM, MHK, IJG, RCP, and MvO) critically reviewed and approved the final draft of the manuscript.

Acknowledgements

The authors thank all participating women in both the MOTHER and follow-up study.

Disclosure statement

Dr B.W. Mol reports grants from NHMRC outside the submitted work. Dr R.C. Painter reports grants from Amsterdam Reproduction and Development, during the conduct of the study, and she is the clinical and scientific advisor to the Dutch patient organization for Hyperemesis Gravidarum. She was also the chair of the organizing committee for the 3rd International Collaboration on Hyperemesis Gravidarum conference, Amsterdam, October 2019. Drs K. Nijsten and all other authors report no conflict of interests. ICMJE disclosure of interest forms are submitted as supporting information.

Data availability statement

The data that support the findings of this study are available from the principle investigator ([email protected]) upon reasonable request.

Additional information

Funding

The MOTHER study was supported by a research grant from the North West Hospital Group, Alkmaar, The Netherlands under grant number 2013T085. The follow-up study was supported by the Amsterdam Reproduction and Development (AR&D) Research Institute, Amsterdam UMC, The Netherlands under project number 23346. Neither funders had any role in the planning, execution or interpretation of this study.