document3

European Journal of Obstetrics & Gynecology and Reproductive Biology xxx (2010) xxx–xxx

G Model

EURO-6842; No of Pages 6

Anxiety and depression in hyperemesis gravidarum: prevalence, risk factorsand correlation with clinical severity

Peng Chiong Tan *, Subramaniam Vani, Boon Kiong Lim, Siti Zawiah Omar

Department of Obstetrics and Gynaecology, Faculty of Medicine, University of Malaya, Lembah Pantai, 50603 Kuala Lumpur, Wilayah Persekutuan, Malaysia

A R T I C L E I N F O

Article history:

Received 14 April 2009

Received in revised form 20 October 2009

Accepted 22 December 2009

Keywords:

Hyperemesis gravidarum

Hospital Anxiety and Depression Scale

Employment

Miscarriage

A B S T R A C T

Objective: To evaluate prevalence, risk factors and clinical severity correlates of anxiety and depression

caseness in hyperemesis gravidarum (HG).

Study design: A prospective study of self-assessment using the Hospital Anxiety and Depression Scale

(HADS) was performed. Women at their first hospitalization for HG were recruited as soon as possible

after hospital admission. Cut-off at the score of 7/8 was used for both the anxiety and depression

subscales of HADS to denote anxiety and depression caseness respectively. Risk factors for anxiety and

depression caseness were identified using Chi-square test, Fisher’s exact test, Mann–Whitney’s U-test or

the Student’s t-test. Multivariable logistic regression analysis incorporating all co-variables with crude

P < 0.1 was performed to identify independent risk factors. Bivariate analyses were performed to

identify associations between clinical markers of severity and anxiety and depression caseness.

Prolonged hospitalization and a number of biochemical and hematological abnormalities were used as

clinical markers of HG severity.

Results: Criteria for anxiety and depression caseness were fulfilled in 98/209 (46.9%) and 100/209

(47.8%) women respectively. 78 (37.3%) participants fulfilled the criteria for both anxiety and depression

caseness, 89 (42.6%) neither, 20 (9.6%) anxiety caseness only and 22 (10.5%) depression caseness only.

Gestational age at commencement of vomiting, duration of vomiting leading up to hospitalization and

paid employment status had crude P < 0.1 in association with anxiety caseness. After adjustment, only

paid employment was independently associated with anxiety caseness (AOR 2.9 95% CI 1.3–6.5;

P = 0.009). Previous miscarriage, gestational age at commencement of vomiting and duration of vomiting

leading up to hospitalization all had P < 0.1 in association with depression caseness. After adjustment,

only previous miscarriage was negatively associated with depression caseness (AOR 0.4 95% CI 0.2–0.9;

P = 0.022). There was no marker of HG severity associated with anxiety caseness on bivariate analysis.

High hematocrit was associated with depression caseness (OR 2.1 95% CI 1.1–3.9; P = 0.027).

Conclusion: Anxiety and depression caseness is common in HG and risk factors can be identified. There is

no convincing association between anxiety and depression and more severe illness. Psychological

symptoms may be a response to physical illness but further studies are needed.

� 2009 Elsevier Ireland Ltd. All rights reserved.

Contents lists available at ScienceDirect

European Journal of Obstetrics & Gynecology andReproductive Biology

journa l homepage: www.e lsev ier .com/ locate /e jogrb

1. Introduction

Hyperemesis gravidarum (HG) occurs in 0.3–2% of pregnancies[1]. HG is characterised by intractable nausea and vomiting due topregnancy, resulting in dehydration and metabolic and biochemi-cal disturbances [2]. HG is empirically differentiated from thecommoner nausea and vomiting of pregnancy (NVP) which affectsup to 80% of pregnancies [3] by the need for hospitalization [2,4].

* Corresponding author. Tel.: +60 3 79492059;

fax: +60 3 79551471.

E-mail addresses: [email protected], [email protected],

[email protected] (P.C. Tan).

Please cite this article in press as: Tan PC, et al. Anxiety and deprecorrelation with clinical severity. Eur J Obstet Gynecol (2010), doi:1

0301-2115/$ – see front matter � 2009 Elsevier Ireland Ltd. All rights reserved.

doi:10.1016/j.ejogrb.2009.12.031

HG is the second ranked indication for hospitalization duringpregnancy in women with live births [5]. Poor maternal weightgain is associated with low birth weight and being small forgestational age [6].

Endocrine associations with HG are often reported, mostconsistently with high levels of human chorionic gonadotropin,but the etiology and pathogenesis of HG remain uncertain and canbe multi-factorial with biologic, psychological and socio-economicantecedents [2].

Historically, nausea and vomiting in pregnancy were thought tohave a psychological basis with various rationales propounded, e.g.arising from resentment or ambivalence towards pregnancy due toimmaturity of personality, strong mother dependence, anxiety andtension related to pregnancy, a sexual disorder resulting from

ssion in hyperemesis gravidarum: prevalence, risk factors and0.1016/j.ejogrb.2009.12.031

mailto:[email protected]



http://dx.doi.org/10.1016/j.ejogrb.2009.12.031

http://www.sciencedirect.com/science/journal/03012115


P.C. Tan et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology xxx (2010) xxx–xxx2

G Model


sexuality aversion, a conversion symptom and a symptom ofhysteria, neurosis or depression, but it is also plausible thatpsychological symptoms are a result of the stress and the physicalburden of HG rather than a cause [2]. HG could also be the result ofpsycho-social stresses, poverty and marital conflicts [2].

Women who terminate their pregnancies secondary to HGoften reported perceived inability to care for the family and self,fear that themselves or their baby could die or that the baby wouldbe abnormal as reasons for their decision, signifying thetremendous physical and psychological burden of the disease tothem [7]. Women with prior psychiatric or medical conditions aremore likely to develop HG when pregnant [8]. In contrast, womenwith HG were no more likely than controls to have psychologicalmorbidity after birth [9].

Data on the prevalence of risk factors of anxiety and depressionstates in clearly defined HG cases are not available. It is also notknown whether psychological distress adversely impact on theclinical severity of HG.

We hypothesise that HG will exacerbate the psychologicalstress of pregnancy. This may result in anxiety and depressionwhich in turn can have an adverse impact on the severity of HG.

We sought to estimate the prevalence and identify of the riskfactors of anxiety and depression caseness in women at firsthospitalization for HG. We also sought to evaluate the relationshipof anxiety and depression caseness with severity of HG as denotedby laboratory and clinical markers.

2. Materials and methods

Women diagnosed with hyperemesis gravidarum and hospi-talized in the University of Malaya Medical Center from 11 Marchto 11 November 2008 were identified and approached on hospitaladmission to enter the study. The patient information sheet wasgiven and queries were answered by the recruiting provider.Patients who agreed to participate gave their written consent. Thestudy was carried out in compliance with the Code of Ethics of theDeclaration of Helsinki. Ethical approval and oversight wereprovided by our institutional review board.

The Hospital Anxiety and Depression Scale (HADS) question-naire comprises 14 stems with four possible answer choices toeach stem. Each answer is scored 0–3. It comprises two subscalesfor anxiety and depression caseness. The HADS has been validatedas a useful tool to identify anxiety and depression cases inantenatal patients and gynecology clinic patients (10). Thesubscale score was summated and cut-off of 7/8 was used forthe summated anxiety and depression scores to differentiatebetween caseness and non-cases [11]. HADS is particularly suitedto investigate psychological distress in pregnancy as it does nothave questions on somatic symptoms [12] – normal pregnancy-related physical symptoms might have inflated the score of otherscales that incorporated somatic symptoms.

The HADS questionnaire and data sheet of personal and socio-economic characteristics were given to participants for self-completion. Oral and written instructions for form filling wereprovided. The forms were collected within 6 h of admission andchecked for completeness. If the questionnaire or data sheet wereincomplete, a further request was made for the forms to becompleted but it was reiterated that a deliberate choice not toprovide specific information would be fully respected.

The HADS questionnaire was reviewed and scored. Any clinicalconcern arising from participants’ responses was followed by adiscussion to ascertain appropriateness for formal psychiatricevaluation. Study data were not routinely made available toproviders.

Inclusion criteria for participation were hospitalization forpresumed HG, a positive urine pregnancy test if a very early


pregnancy was not visible on ultrasound or a viable singleton fetuson ultrasound with gestational age < 14 weeks. Women withrecurrent admissions for HG to our center were only recruitedduring their first hospitalization. Women were excluded if theyhad thyroid disease, gestational trophoblastic disease, psychiatricillness or any other acute confounding illness that could causenausea or vomiting.

Laboratory results and clinical data were abstracted onto the casereport form from case notes and computerized laboratory records.Ketonuria [13], renal function and serum electrolytes [14] and fullblood count [15] results were used as measures of severity of HG onhospitalization. These investigations were performed before sub-stantive treatment had started. Ketonuria of 3+ to 4+ on urinedipstick was classified as severe. Cut-offs for high urea, creatinineand hematocrit were set to demarcate the upper quartile in ourstudy population for these parameters as pregnancy significantlylowered their range. The laboratory-supplied normal range wasused to demarcate low serum sodium, low serum potassium, anemiaand leucocytosis. The length of hospital stay was also used as ameasure of disease severity: 4 days’ stay was categorised asprolonged as this cut-off demarcated the upper quartile.

Standard inpatient management of HG was given to partici-pants. All were initially rehydrated with intravenous fluids, andgiven intravenous anti-emetics and oral thiamine. Oral intake wasresumed based on clinical judgment. Patients were dischargedonce they were rehydrated and capable of maintaining adequateoral intake.

About 50% of women with nausea and vomiting in pregnancyhave been reported to have potential psychiatric problems [16]. Ina recent report from our center, about 25% required prolonged (�4days) hospitalization at their first admission for HG [15]. Samplesize calculation using prolonged hospitalization as the primemarker for clinical severity, a 0.05, power 0.8, 1 to 1 ratio ofpsychological caseness to non-caseness, applying Fisher’s exacttest and factoring in an absolute 20% increase from 25% to 45% forprolonged hospitalization, indicated that 196 women in total wererequired for a powered study.

Statistical analysis was performed using the SPSS 15 (SPSS Inc.,Chicago, IL, USA). The one sample Kolmogorov–Smirnoff test wasused to assess normality of data distribution. Fisher’s exact testwas used for 2 � 2 categorical data, Chi-square test for larger than2 � 2 categorical data sets. The means of normally distributedcontinuous data were assessed by Student’s t-test. The Mann–Whitney’s U-test was used for non-normally distributed data andordinal data. Multivariable logistic regression analysis was usedto control for co-variables with P � 0.1 on bivariate analysis.P < 0.05 on two-sided tests was taken as a level of significance.

3. Results

There were a total of 284 admissions to our center for HG overthe study period of March–November 2008. The recruitment flowchart (Fig. 1) depicted the movement of all potential participantsand their removal from study participation due to criteriaexclusions, failure to recruit and refusal to participate. 209/215(97%) of eligible patients were recruited and analysed. No womanwas excluded because of a previous psychiatric history.

98/209 (46.9%) of participants scored �8 on the anxietycomponent of the HADS and fulfilled the criterion for anxietycaseness. 100/209 (47.8%) similarly fulfilled criterion for depres-sion caseness.

There was a strong association and overlap between casenessfor anxiety and depression caseness (OR 15.8 95% CI 8.0–31.1;P < 0.001). 78 (37.3%) participants fulfilled the criteria for bothanxiety and depression caseness, 89 (42.6%) neither, 20 (9.6%)anxiety caseness only and 22 (10.5%) depression caseness only.



Fig. 1. Recruitment flow chart of hyperemesis gravidarum patients into study.

Table 1Risk factors for high anxiety score (�8) using the Hospital Anxiety and Depression Sca

High anxiety

score (�8) n = 98

Norm

score

Age 28.2�4.3 28.0�

Ethnicity

Malay 74 (75.5) 78 (70

Chinese 7 (7.1) 6 (5.4

Indian 15 (15.3) 19 (17

Others 2 (2.0) 8 (7.2

Gravidity 2 [1] 2 [2]

Parity 0 [1] 1 [2]

Nulliparous 55 (56.1) 54 (48

Previous miscarriage 15 (15.3) 24 (21

Gestational age at admission (weeks) 9.0�2.3 9.1�2

Gestational age at start of vomiting (weeks) 6.4�2.5 6.9�2

Duration of vomiting at admission (days) 14 [21] 14 [14

Frequency of vomiting (per day) 7 [5] 8 [5]

Hyperemesis in a previous pregnancy 34 (34.7) 48 (43

Medical illnessb 13 (13.3) 15 (13

Planned pregnancy 45 (45.9) 45 (40

Previous use of combined contraceptive pill 16 (16.3) 21 (18

Married 95 (96.9) 109 (9

Local family support 64 (65.3) 67 (60

Incomec

Higher income (>RM3000 per month) 49 (50.5) 55 (49

Low income (�RM 3000 per month) 48 (49.5) 56 (50

Housing

Owned 37 (37.8) 40 (36

Rented 52 (53.1) 60 (54

Living in extended family 9 (9.2) 11 (9.

Educational attainment

Tertiary 59 (60.2) 57 (51

Up to secondary 39 (39.8) 54 (48

Employment

Paid 88 (89.8) 85 (76

Otherd 10 (10.2) 26 (23

Regular exercise before pregnancy 34 (34.7) 41 (36

Values expressed as mean� standard deviation, number (%) or median [interquartile rang

Mann–Whitney’s U-test for ordinal data and non-normally distributed continuous data an

categorical data. Multivariable logistic regression analysis model incorporated all variablea Adjusted odds ratio AOR (95% Confidence Interval) shown where adjusted P<0.05b Medical illness included cases with mainly asthma (15 women), chronic hypertensc Family income: US $1 = RM3.5.d Included housewives (32 women) and students (4).

P.C. Tan et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology xxx (2010) xxx–xxx 3

G Model



Results of bivariate analyses performed for the recordedcharacteristics stratified according to anxiety caseness or depres-sion caseness followed by multivariable logistic regressionanalyses incorporating and controlling for all co-variables withcrude P < 0.1 were shown in Tables 1 and 2 respectively.

On bivariate analyses for anxiety caseness, gestational age atcommencement of vomiting, duration of vomiting leading up tohospitalization and paid employment status of the participants allhad P < 0.1. After adjustment, only paid employment wasidentified as an independent risk factor for anxiety caseness(AOR 2.9 95% CI 1.3–6.5; P = 0.009).

For depression caseness, previous miscarriage, gestational ageat commencement of vomiting and duration of vomiting leadingup to hospitalization all had P < 0.1 on bivariate analyses.Following adjustment, only previous miscarriage had an indepen-dent negative association with depression caseness (AOR 0.4 95%CI 0.2–0.9; P = 0.022).

Table 3 shows the results of the bivariate analyses performedseparately for anxiety and depression caseness against a number ofclinical, metabolic and laboratory markers of severity of HG. Therewas no marker of HG severity against anxiety caseness on bivariateanalysis with a P < 0.05; only hyponatremia (OR 0.6 95% CI 0.3–1.0; P < 0.073) was borderline for a negative association with

le in hyperemesis gravidarum.

al anxiety

(�7) n = 111

P value AOR

(95% CI)a

Adjusted

P value

4.7 0.86

.3) 0.33

)

.1)

)

0.13

0.19

.6) 0.28

.6) 0.29

.3 0.67

.3 0.08 0.27

] 0.06 0.31

0.24

.2) 0.26

.5) 1.0

.5) 0.49

.9) 0.72

8.2) 0.67

.4) 0.48

.5) 1.0

.5)

.0) 0.96

.1)

9)

.4) 0.21

.6)

.6) 0.016 AOR 2.9 (95% CI 1.3–6.5) 0.009

.4)

.9) 0.78

e]. Student’s t-test was used to analyse continuous data that is normally distributed,

d Fisher’s exact test (2�2 datasets) or Chi-square test (larger than 2�2 datasets) for

s with crude P� 0.1 on bivariate analysis.

.

ion (4), diabetes (3), combined hypertension diabetes (1) and others (5).



Table 2Risk factors for high depression score (�8) using the Hospital Anxiety and Depression Scale in hyperemesis gravidarum.

High depression

score (�8) n = 100

Normal depression

score (�7) n = 109

P value AOR (95% CI)a Adjusted

P value

Age 27.9�4.3 28.2�4.7 0.66

Ethnicity

Malay 78 (78) 74 (67.9) 0.38

Chinese 5 (5.0) 8 (7.3)

Indian 14 (14.0) 20 (18.3)

Others 3 (3.0) 7 (6.4)

Gravidity 1.5 [1] 2 [2] 0.12

Parity 0 [1] 0 [2] 0.51

Nulliparous 53 (53.0) 56 (51.4) 0.89

Previous miscarriage 12 (12.0) 27 (24.8) 0.021 AOR 0.4 (95% CI 0.2–0.9) 0.022

Gestational age at admission (weeks) 9.0�2.3 9.1�2.3 0.78

Gestational age at start of vomiting (weeks) 6.3�2.6 7.0�2.2 0.043 0.19

Duration of vomiting at admission (days) 14 [21] 14 [10.5] 0.011 0.32

Frequency of vomiting (per day) 7 [5] 8 [5] 0.63

Hyperemesis in a previous pregnancy 35 (35.0) 47 (43.1) 0.26

Medical illnessb 12 (12.0) 16 (14.7) 0.69

Planned pregnancy 45 (45.0) 45 (41.3) 0.68

Previous use of combined contraceptive pill 14 (14.0) 23 (21.1) 0.21

Married 97 (97.0) 107 (98.2) 0.67

Local family support 65 (65.0) 66 (60.6) 0.57

Incomec

Higher income (>RM3000 per month) 52 (52.5) 52 (47.5) 0.58

Low income (�RM3000 per month) 47 (47.5) 57 (52.3)

Housing

Owned 35 (35.0) 42 (38.5) 0.87

Rented 55 (55.0) 57 (52.3)

Living in extended family 10 (10) 10 (9.2)

Educational attainment

Tertiary 60 (60.0) 56 (51.4) 0.27

Up to secondary 40 (40.0) 53 (48.6)

Employment

Paid 87 (87.0) 86 (78.9) 0.144

Otherd 13 (13.0) 23 (21.1)

Regular exercise before pregnancy 38 (38.0) 37 (33.9) 0.57

Values expressed as mean� standard deviation, number (%) or median [interquartile range]. Student’s t-test was used to analyse continuous data that is normally distributed,

Mann–Whitney’s U-test for ordinal data and non-normally distributed continuous data and Fisher’s exact test (2�2 datasets) or Chi-square test (larger than 2�2 datasets) for

categorical data. Multivariable logistic regression analysis model incorporated all variables with crude P�0.1 on bivariate analysis.a Adjusted odds ratio AOR (95% Confidence Interval) shown where adjusted P<0.05.b Medical illness included cases with mainly asthma (15 women), chronic hypertension (4), diabetes (3), combined hypertension diabetes (1) and others (5).c Family income: US $1 = RM3.5.d Housewives (32 women) and students (4).


G Model


anxiety caseness. High hematocrit was associated with depressioncaseness (OR 2.1 95% CI 1.1–3.9; P = 0.027).

We approached participants who fulfilled caseness criteria foranxiety or depression with an offer for a formal psychiatricappointment. The offer was not taken up by any woman.

Table 3Laboratory and clinical markers of the severity of hyperemesis gravidarum and high a

High anxiety

score (�8) n = 98

Normal anxiety

score (�7) n = 111

Prolonged hospitalization (�4 days)a 40 (40.8) 35 (31.5)

Ketonuria

3+ to 4+ 66 (67.3) 80 (72.1)

1+ to 2+ 32 (32.7) 31 (27.9)

Hyponatremia (�135 mmol/l)b 46 (46.9) 66 (59.5)

Hypokalemia (�3.5 mmol/l)b 20 (20.4) 22 (19.8)

High urea (�3.3 mmol/l)a 21 (21.4) 30 (27.0)

High creatinine (�55 mmol/l)a 31 (31.6) 24 (21.6)

Anemia (hemoglobin <11.5 g/dL)b 11 (11.2) 14 (12.6)

High hematocrit (�0.41)a 30 (30.6) 24 (21.6)

Leucocytosis (>11�109/l)b 35 (35.7) 35 (31.5)

Values expressed as number (%). Fisher’s exact test (2�2 datasets) used for analysis. Mu

crude P<0.1 on bivariate analyses.a Cut-offs demarcating the upper quartile of values in our study population.b Cut-off values supplied by our hospital laboratory.


4. Comments

Anxiety and depression caseness were common in womenaffected by HG – more than a third of our participants fulfilledcriteria for both anxiety and depression caseness. Almost half the

nxiety or depression score (�8) using the Hospital Anxiety and Depression Scale.

P value High depression

score (�8) n = 100

Normal depression

score (�7) n = 109

P value

0.19 40 (40.0) 35 (32.1) 0.25

0.55 68 (68.0) 78 (71.6) 0.65

32 (32.0) 31 (28.4)

0.073 51 (51.0) 61 (56.0) 0.49

1.0 23 (23.0) 19 (17.4) 0.39

0.42 23 (23.0) 28 (25.7) 0.75

0.12 31 (31.0) 24 (22.0) 0.16

0.83 8 (8.0) 17 (15.6) 0.13

0.16 33 (33.0) 21 (19.3) 0.027

0.56 38 (38.0) 32 (29.4) 0.19

ltivariable logistic regression analysis was planned incorporating co-variables with



P.C. Tan et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology xxx (2010) xxx–xxx 5

G Model


participants were potential cases for anxiety and a similarproportion for depression caseness. These prevalence rates werehigher than that found in unselected antenatal population fromHong Kong who were also assessed with HADS in the first trimesterof pregnancy. Their anxiety caseness prevalence rate was 36.3%,depression caseness rate 22.1% and 14.2% had both anxiety anddepression caseness [12]. A recent Scottish study on anxiety anddepression after miscarriage using HADS but with cut-off set at 10/11 shows prevalence rates for anxiety caseness of 28.3% and fordepression caseness of 10% 1 month after miscarriage [17].Reconfiguring our data to reflect HADS subscale cut-off at 10/11,prevalence rates for our HG women were 20.6% and 29.2%respectively for anxiety and depression caseness. This suggeststhat women hospitalized for HG experienced psychologicaldistress of a similar magnitude to women with a recentmiscarriage. We were unable to identify other studies of anxietyand depression in clearly defined HG cases to compare ourfindings.

Paid employment was found to be an independent risk factorfor anxiety in our participants. The anxiety felt by women in paidemployment seemed to be well placed and perceptive. A recentsurvey has indicated that 82.8% of women with HG reportednegative psycho-social changes: socio-economic consequencesinvolving job loss being an important burden [18]. The medianduration of symptoms in our study population was 14 days[interquartile range 14.5] and mean gestational age at hospitali-zation was 9 weeks. Employees in our country are typicallyallowed 14 days ‘‘sick leave’’ per year before potentially facing payreductions. It is well known that vomiting is an early pregnancyproblem that resolves in 90% of cases by the 16th week of gestation[3]. It is conceivable that even a relatively short period of illness ofabout 2 weeks coupled with the perception that they might notcompletely recover for another few weeks could have raised theanxiety level of women who were in paid employment. We did notpose questions about how much time the women had alreadytaken off work by the time of hospitalization or employmentconcerns in our study design. No woman in our study populationwas ‘‘unemployed’’. Those classified as not in paid employmentwere housewives or students. Hence our finding should beappropriately interpreted.

Previous miscarriage was the only independent ‘‘protective’’factor against depression caseness. Nausea and vomiting inpregnancy have long been known to be strongly associated withreduced risk of miscarriage in early pregnancy as reported by ameta-analysis from 1989 [19]. Vomiting in pregnancy may evenbe associated with reduced miscarriage risk in threatenedmiscarriage [20]. However, we did not ascertain whetherknowledge about this protective association was widespread inour study population. Awareness that vomiting in pregnancy isassociated with reduced risk of miscarriage could account for ourfinding of previous miscarriage being a protective factor againstdepression caseness.

Of the many markers of laboratory and clinical severity of HGanalysed against anxiety and depression caseness, only a highhematocrit on hospitalization was positively associated withdepression caseness. No other co-variables reached P < 0.1 againstdepression caseness. High hematocrit in HG could be secondary tosevere dehydration and it has been independently associated withprolonged hospitalization [15]. However, depression caseness wasnot significantly associated with prolonged hospitalization in ouranalysis. There was no significant association of any of the severitymarkers considered with anxiety caseness. There was a borderlinenegative association between hyponatremia and anxiety caseness(P = 0.07). Given the multiple analyses performed, the isolatedassociation with high hematocrit should be interpreted withcaution. Overall there is no convincing pattern of anxiety and


depression caseness being more associated with more severe HGsince many other markers were tested besides hematocrit.

A recent report on NVP has concluded that there is an associationbetween anxiety and depression early in pregnancy and severity ofnausea and vomiting [21]. Taken in their entirety, our data did notsupport an association between psychological caseness in HG and amore severe physical illness as characterised by metabolic andbiochemical disturbances at presentation or by prolonged hospital-ization. Duration of vomiting prior to hospitalization and frequencyof vomiting were also not independently associated with psycho-logical caseness. Our findings in women with HG seemed to bedifferent from the aforementioned study on women with NVP.

Previous reports evaluating psychology in the etiology andpathogenesis of HG have been conflicting. While a psychiatrichistory prior to HG is reportedly uncommon, a psychiatric historymay increase the risk of HG [8], but psychological disordercommonly resolves simultaneously with physical symptoms [9].We did not have to exclude any HG cases from the study based onpre-established criteria of prior psychiatric illness, implying thatwomen admitted to our institution generally did not have priormajor psychological issues. Also when women were identified asfulfilling anxiety or depression caseness, none took up the offer offormal psychological assessment. None of these women meritedcompulsory psychological assessment based on normal clinicalpractice. We do not think these women lacked insight – we believethey declined the offer of formal assessment because they thoughtthat any psychological disturbance they experienced was due tothe physical illness of HG and should resolve once the nausea andvomiting remitted. Women with HG viewed HG as biologicallydetermined [22]. Our finding that psychological caseness was notassociated with more severe HG, and the independent risk factorsidentified, would be more consistent with psychological distressbeing a response to the physical illness rather than the driver in thepathogenesis of HG.

There were other limitations to our study. HADS has not beenvalidated in HG though it is validated for the pregnant population[10]. Longitudinal as well as control data on psychological distressprior to pregnancy and after recovery from HG were not availableso the mechanism whereby psychological distress led to HG or HGcontributed to subsequent psychological consequences cannot beconclusively evaluated with our data. 40.8% of women withanxiety caseness required prolonged hospitalization compared to31.5% of non-anxious controls. Post hoc analysis indicated that ourstudy had 24% power only to detect the above stated difference inthe rate of prolonged hospitalization, which would suggest that theassumption made in our sample size calculation was optimistic.

There is a heavy burden of psychological distress in women withHG. It is unclear how these stresses evolve over the course of theillness and pregnancy. A powered prospective longitudinal studystarting from before pregnancy and followed into the postnatalperiod in women at high risk of HG will help provide quality data onthe cause and effect between HG and psychological states. It is notestablished whether support and therapy aimed at anxiety anddepression would be useful avenues to better outcome. However, alack of health care provider support and appreciation of diseaseseverity was more often reported by women with HG whounderwent termination of pregnancy [7]. Also patients’ perceptionsabout physician humanism and perceived agreement about thecause and impact of HG can help patient satisfaction [23]. Furtherresearch in this direction is warranted.

5. Conclusion

Anxiety and depression caseness is common in hyperemesisgravidarum. Independent risk factors for psychological distress canbe identified. There is no convincing association between anxiety




G Model


and depression and more severe illness. Psychological symptoms may

be a response to physical illness but further studies are needed.

References

[1] Ismail SK, Kenny L. Review on hyperemesis gravidarum. Best Pract Res ClinGastroenterol 2007;21(5):755–69.

[2] Verberg MF, Gillott DJ, Al-Fardan N, Grudzinskas JG. Hyperemesis gravidarum, aliterature review. Hum Reprod Update 2005;11(September–October (5)):527–39.

[3] Gadsby R, Barnie-Adshead AM, Jagger C. A prospective study of nausea andvomiting during pregnancy. Br J Gen Pract 1993;43(June (371)):245–8.

[4] Bailit JL. Hyperemesis gravidarium: epidemiologic findings from a large co-hort. Am J Obstet Gynecol 2005;193(September (3 Pt 1)):811–4.

[5] Gazmararian JA, Petersen R, Jamieson DJ, et al. Hospitalizations during pregnancyamong managed care enrollees. Obstet Gynecol 2002;100(July (1)):94–100.

[6] Dodds L, Fell DB, Joseph KS, Allen VM, Butler B. Outcomes of pregnanciescomplicated by hyperemesis gravidarum. Obstet Gynecol 2006;107(February(2 Pt 1)):285–92.

[7] Poursharif B, Korst LM, Macgibbon KW, Fejzo MS, Romero R, Goodwin TM.Elective pregnancy termination in a large cohort of women with hyperemesisgravidarum. Contraception 2007;76(February (6)):451–5.

[8] Seng JS, Schrot JA, van De Ven C, Liberzon I. Service use data analysis of pre-pregnancy psychiatric and somatic diagnoses in women with hyperemesisgravidarum. J Psychosom Obstet Gynaecol 2007;28(December (4)):209–17.

[9] Simpson SW, Goodwin TM, Robins SB, et al. Psychological factors andhyperemesis gravidarum. J Womens Health Gend Based Med 2001;10(June(5)):471–7.

[10] Abiodun OA. A validity study of the Hospital Anxiety and Depression Scale ingeneral hospital units and a community sample in Nigeria. Br J Psychiatry1994;165(November (5)):669–72.

[11] Zigmond AS, Snaith RP. The hospital anxiety and depression scale. ActaPsychiatr Scand 1983;67(June (6)):361–70.


[12] Lee AM, Lam SK, Sze Mun Lau SM, Chong CS, Chui HW, Fong DY. Prevalence,course, and risk factors for antenatal anxiety and depression. Obstet Gynecol2007;110(November (5)):1102–12.

[13] Tan PC, Jacob R, Quek KF, Omar SZ. The fetal sex ratio and metabolic,biochemical, haematological and clinical indicators of severity of hyperemesisgravidarum. BJOG 2006;113(June (6)):733–7.

[14] Tan PC, Jacob R, Quek KF, Omar SZ. Pregnancy outcome in hyperemesisgravidarum and the effect of laboratory clinical indicators of hyperemesisseverity. J Obstet Gynaecol Res 2007;33(August (4)):457–64.

[15] Tan PC, Jacob R, Quek KF, Omar SZ. Indicators of prolonged hospital stay inhyperemesis gravidarum. Int J Gynaecol Obstet 2006;93(June (3)):246–7.

[16] Swallow BL, Lindow SW, Masson EA, Hay DM. Psychological health in earlypregnancy: relationship with nausea and vomiting. J Obstet Gynaecol 2004;24(January (1)):28–32.

[17] Cumming GP, Klein S, Bolsover D, et al. The emotional burden of miscarriagefor women and their partners: trajectories of anxiety and depression over 13months. BJOG 2007;114(January (9)):1138–45.

[18] Poursharif B, Korst LM, Fejzo MS, MacGibbon KW, Romero R, Goodwin TM. Thepsychosocial burden of hyperemesis gravidarum. J Perinatol 2008;March(28):176–81.

[19] Weigel RM, Weigel MM. Nausea and vomiting of early pregnancy and preg-nancy outcome. A meta-analytical review. Br J Obstet Gynaecol 1989;96(March (11)):1312–8.

[20] Weigel MM, Weigel RM. Nausea and vomiting of early pregnancy and preg-nancy outcome. An epidemiological study. Br J Obstet Gynaecol 1989;96(No-vember (11)):1304–11.

[21] Koken G, Yilmazer M, Cosar E, Sahin FK, Cevrioglu S, Gecici O. Nausea andvomiting in early pregnancy: relationship with anxiety and depression. JPsychosom Obstet Gynaecol 2008;29(June (2)):91–5.

[22] Munch S. Women’s experiences with a pregnancy complication: causal expla-nations of hyperemesis gravidarum. Soc Work Health Care 2002;36(1):59–76.

[23] Munch S, Schmitz MF. Hyperemesis gravidarum and patient satisfaction: apath model of patients’ perceptions of the patient–physician relationship. JPsychosom Obstet Gynaecol 2006;27(March (1)):49–57.



document3

Documents