Pruritus in pregnancy: a study of anatomical distribution and prevalence in relation to the development of obstetric cholestasis

Abstract

The main objective of this study was to determine the prevalence and anatomical distribution of pruritus in 6532 pregnant women from a UK antenatal population. Pregnant women attending and completing antenatal care at two general hospitals over a 12-month period were recruited and contacted on three occasions by post. Medical advice and a questionnaire detailing the nature and severity of their pruritus were included. Pruritus in pregnancy, as reported by questionnaire, affected approximately 23% of pregnancies (n = 1521/6532 women) and 1.6% (n = 25) of these women developed obstetric cholestasis (OC). Overall, 0.66% of the antenatal population (43/6532) had a clinical diagnosis of OC (95% CI: 0.48–0.89%). Itching unrelated to OC was most commonly reported to be worst on the abdomen (31%, 616/2014). Women with OC reported pruritus to be most severe on the palms and soles in 16% (4/25) and ‘all over’ in 24% (6/25) compared with 5% (54/1120) (P < 0.05) and 4% (42/1120, P < 0.0001) of those without OC. In conclusion, pruritus affected approximately one in four women and OC one in 135 women during the study period. Women whose pruritus is ‘all over’ or most severe on the ‘palms or soles’ may be at greater risk of the disease.

INTRODUCTION

Anecdotally, pruritus in pregnancy is perceived to be common. A potential cause of pruritus in pregnancy (in the absence of rash) is obstetric cholestasis (OC), which is defined as pruritus with abnormal liver function in the absence of other liver pathology, and which resolves following delivery. As OC is associated with adverse maternal and fetal outcomes including intrauterine death, premature labour, meconium stained liquor and postpartum haemorrhage,^1–4 many women with pruritus are investigated. A more precise understanding of the clinical nature and presentation of pruritus and OC in the UK antenatal population could reduce the number of unnecessary investigations in women with benign pruritus.

A high incidence of OC has been reported in Scandinavia and Chile (2% and 4%, respectively),^5,6 but an accurate prospective prevalence in a UK population has never been determined. Similarly, there are only a few descriptions of pruritus in pregnancy outside the context of OC and these have concentrated on highly selected populations from tertiary referrals to dermatology and other specialist clinics. Earlier reports also used varying definitions for the pathology observed and differing inclusion and exclusion criteria, e.g. excluding or only including those with a rash.^7–10 Extrapolation of these data to the general obstetric population is therefore impossible. There are also no previous studies that have systematically prospectively screened a general antenatal population for pruritus in pregnancy or attempted to describe the pruritus reported in relation to development of OC. In this prospective study, we have assessed the incidence and described the anatomical distribution of pruritus associated with benign causes and that associated with OC in a large UK antenatal cohort.

METHODS

The study strategy was to identify all women with pruritus and/or OC that attended for antenatal care at two hospitals serving the general population of two hospitals (West Middlesex University Hospital and Whipps Cross University Hospital) over a 12-month period starting in September 1999. This was achieved in two ways. (i) All women were contacted (with ethical approval at each site) by letter on four occasions (early pregnancy, 30 and 36 weeks' of pregnancy and following delivery), which explained the importance of reporting pruritus in pregnancy in order that liver function tests and serum bile acid determination could be performed to exclude OC. A simple questionnaire which women with pruritus were invited to complete was also included with a freepost reply envelope. Posters with similar information, additional questionnaires, translations and ‘post boxes’ for replies were made available on antenatal wards, clinics and ultrasound departments at both sites. An additional questionnaire for completion following delivery was also provided. Questions were asked relating to any association with a rash, the location of pruritus, any personal or family history of pruritus or atopy and the gestation at onset of symptoms. The severity of pruritus was measured using a modified scoring system (1 = mild to 5 = severe).^11,12 (ii) Staff at the participating hospitals were made aware of the study via posters and lectures to ensure a high level of vigilance and comprehensive screening of all women with pruritus and/or suspected OC. Staff were asked to report all identified women to the study team and/or the dedicated referral clinic.

The diagnosis of OC, i.e. pruritus and abnormal liver function tests was further standardized and based upon (i) persistent pruritus of onset in pregnancy in the absence of a rash other than excoriations; (ii) abnormal liver function tests (one or more of abnormal gamma glutamyl-transpeptidase, alanine amino-transferase, aspartate amino-transferase, total bile acids); (iii) the absence of other relevant liver disease identifiable on testing antismooth muscle antibodies, antimitochondrial antibodies, Epstein–Barr virus, cytomegalovirus, hepatitis A, hepatitis B, hepatitis C and a liver ultrasound; and (iv) postnatal normalization of serum biochemistry and cessation of symptoms. All women with OC were reviewed by a member of the study team to confirm the diagnosis. This included some women who had chosen not to complete questionnaires. In any case of OC those who had not participated in the additional questionnaire-based study previously were encouraged not to do so in order to avoid any bias introduced as a result of questioning during the clinical consultation. In those with a pruritic rash where the diagnosis was in doubt, a consensus was reached by more than one clinician, including a dermatologist when necessary. Polymorphic eruption of pregnancy (PEP) was defined as pruritic, urticarial papules and plaques on the abdomen along striae, with periumblical sparing and presenting in the second or third trimester. Prurigo was defined as pruritic red/brown excoriated papules on extensor surface of limbs. Pruritic folliculitis was defined as acneiform pruritic erythematous, follicular papules and pustules on the trunk and/or thighs in the second or third trimester. Suspected cases of pemphigoid gestationes (PG) (vesicles or bullae on the limbs, palms, soles or abdomen with umbilical involvement) were referred for dermatology opinion in all cases.^13,14 Skin biopsy was performed when a clinical diagnosis was not possible and in all cases of PG. Those women with pruritus and normal liver function in the absence of a rash or identifiable dermatological condition were classified as pruritus gravidarum. Serial liver function tests were performed in women whose symptoms persisted, and these data have been published previously.¹⁵

Data analysis

The total number of women attending for antenatal care was obtained from hospital records. Overall prevalence of pruritus and OC were determined from the questionnaire responses and clinical information collated by the study team. Questionnaire data were also assessed independently. Returned questionnaires were coded and data entered on a Microsoft EXCEL 97 spreadsheet and imported to Stata version 6.0 (StataCorp, College Station, TX, USA) for analysis. Percentages were compared using the χ ² test, with P < 0.05 taken to represent statistical significance. When a woman returned more than one questionnaire, data representing the most advanced gestation were analysed.

RESULTS

During the study period at the two hospitals, 7917 women were referred for pregnancy care, of whom 6532 completed their pregnancies, 0.66% of these women were diagnosed with OC (43/6532, 95% CI: 0.48–0.89%).

In total, 1521 (23%) women with completed pregnancies returned questionnaires including 25 (1.6%) women with OC.

Data provided by women with pruritus indicated that 40% (610/1521) noted pruritus in association with a rash. The most commonly reported site of this rash was the abdomen (45%; 277/610).

Most women (99%; 1504/1521) reported new onset of pruritus in the current pregnancy and only 1% prior to pregnancy. An allergy (e.g. food) or infection (e.g. fungal infection) was the attributed cause in 3% (44/1436) and 1% (19/1436), respectively, of those with pruritus.

The largest proportion of women (95/1521) perceived pruritus to occur between 9 and 12 weeks gestation (Figure 1). Pruritus was most often mild (scoring 1/5) (22%, 338/1521) and severe in only 2% (35/1521). The most frequently reported time of day at which the pruritus was perceived to be most severe was the evening (21%, 318/1521) or at night (19%, 293/1521).

Figure 1.

Gestation at the onset of pruritus: as recalled by 1521 women completing questionnaires

When asked to name a single site where the pruritus was most severe, the abdomen was highly represented (41%, 455/1120; Figure 2). However, women with OC were more likely to report severe pruritus ‘all over’ (24%) and on ‘the legs’ (28%; Figure 2) compared with the 4% of women with severe pruritus who did not develop OC (P < 0.0001). Only three cases (12%) of those with OC reported pruritus to be most severe on their abdomen. In those women with pruritus in pregnancy who did not develop OC, 5% reported the palms and soles to be the sites of most severe pruritus compared with 16% in the OC population (P < 0.05).

Figure 2.

Anatomical location reported to be the site of most severe pruritus in women with obstetric cholestasis compared with those with pruritus and normal liver function. Black bars represent those women with pruritus. Grey bars represent those with obstetric cholestasis

Four hundred and eighty-one women with pruritus in pregnancy completed postnatal questionnaires and 76% (367/481) of these reported that pruritus had ceased postnatally with 40% (114/284) specifying that pruritus had ceased following delivery, the vast majority (86%) stipulating it to be immediately after birth (98/114).

Additional suggested risk factors for pruritus in pregnancy such as a personal (or family) history of atopy, pruritus when taking the ‘pill’ or cyclically were analysed. Fourteen percent (213/1521) of women with pruritus in the current pregnancy had a parous female family member who also complained of pruritus in pregnancy. Thirty-six percent (539/1521) reported a personal, and 52% (786/1521) a family history of atopy (asthma, eczema or hayfever). Eight percent of women (119/1521) had noted pruritus in relation to their menstrual cycle or use of the contraceptive pill.

Of the multiparous women with pruritus in the current pregnancy, 79% (439/556) had itched in a previous pregnancy.

Within the study population, two hundred women with pruritus were referred to, or contacted, the study team directly. Diagnoses among these women are shown in Figures 3 and 4.

Figure 3.

Diagnoses among women with pruritus who were reviewed by the study team. Two hundred women were referred for review by the study team. The clinical diagnosis concluded as a result of that meeting is shown. PG = pruritus gravidarum; OC = obstetric cholestasis; PPP = pruritus preceding pregnancy; TALFT = transiently abnormal liver function tests

Figure 4.

Final diagnoses among women with a pruritic rash. Thirty-nine women with a rash had a dermatological cause for their pruritic rash

DISCUSSION

This, to our knowledge, is the first study to have prospectively characterized the anatomical locations most likely to be associated with benign pruritus and the pruritus associated with OC. These data provide potentially clinically useful information that may assist in the differentiation of benign pruritus of pregnancy from that reported in OC. In addition, we have determined the prevalence of pruritus and OC in a UK population. In the antenatal study population targeted for information regarding pruritus in pregnancy, approximately one in four pregnancies were affected by pruritus, and 0.66% of them developed OC (1 in 135 pregnancies). This incidence is similar to that reported in a retrospective UK study in 1999 (0.7%).⁵

After exclusion of previous pruritus, infection or allergy, 21% of the antenatal population reported pruritus. This is much higher than has been reported previously (1.2–18%),^{7–10,16–18} and is likely to reflect our systematic approach to contacting all pregnant women and regular surveillance of antenatal clinics. Indeed, this figure may be an underestimation as our questionnaire-based data will have inevitably excluded some non-responders who developed pruritus. Overall, this study strongly confirms the widely held belief among clinicians that pruritus is a common feature of pregnancy and is more common than previously shown.

It is also anecdotally reported by some clinicians that pruritus in pregnancy is more often associated with advanced gestation. However, in the population studied, the onset of pruritus predominantly occurred in early or mid gestation. The former generalizations probably originate from literature related to the dermatoses of pregnancy which can often present later in pregnancy.¹⁹ Indeed, rapid abdominal wall distension, associated with later gestations, is thought to result in connective tissue damage and an inflammatory response leading to PEP.²⁰

The data from the present study suggest that in many cases pruritus in pregnancy is mild, but not unsurprisingly it is often perceived to be more severe in the evening and at night. At these times relaxation and reduction in other exogenous stimuli may lead to an altered perception of the severity of pruritus.

Many of the women (36%) who reported pruritus in pregnancy had a personal history of atopy and 44% reported a rash. Eczema may be more common in pregnancy, affecting up to 36% of pregnant women in dermatology clinics.¹³ In an atopic individual, eczema may flare premenstrually or present for the first time in pregnancy.²¹ Female sex steroids have been implicated in the pathogenesis of this type of eczema.²¹ The mechanical action of scratching in an atopic woman can itself lead to further skin abnormalities and this is a suggested mechanism in prurigo.²² A recent suggestion has been to group eczema in pregnancy, prurigo and pruritus folliculitis together as ‘atopic eruption of pregnancy’ to reflect their common aetiology.²³ While we have not shown any correlation between cyclical pruritus and pruritus in pregnancy, new onset of eczema in pregnancy, or a rash resulting from scratching in an atopic woman, may have been a possible cause in 15% of the cases of pruritus in our survey. A permissive role for female sex steroids may be supported by the observations that 40% of pruritus ceased postnatally.

The classification of the various forms of pruritus unique to pregnancy has varied widely over many years. Many of the clinical definitions overlap and refer to similar conditions, and therefore studies attempting to define prevalence maybe misleading. Agreed classifications are essential, and prevalence statistics are helpful when managing the pregnant woman with pruritus. Investigation and referral to a specialist may then be focused on those likely to have pathology.

Of those women in the study population reporting a rash of new onset in pregnancy, a number were reviewed by the study team and one of the pregnancy-specific dermatoses (PEP, pruritic folliculitis, prurigo and PG) was diagnosed. Serological and histological diagnosis was not possible in every case. However, the incidence of the dermatoses of pregnancy in the subgroup reported in this study is compatible (where known) with that observed in the general population (i.e. PEP 1:200–250, pruritic folliculitis not known, prurigo 1:300 and PG 1:60,000).¹³

The questionnaire may have elicited a falsely high prevalence for the presence of a rash. The presence of a rash is subjective, with some pregnant women mistaking excoriations for a dermatosis. For example, 10/25 women diagnosed with OC reported a rash on completing the questionnaire but this was not defined as such on clinical examination.

Although an important condition, our data clearly show that OC represents only a small proportion of the morbidity resulting from pruritus in pregnancy. Effective treatment of pruritus is not available in all cases. Treatments for the specific dermatoses of pregnancy (commonly topical steroids) are often effective, but for those women with pruritus gravidarum the treatment is largely empirical with emollients and antihistamines. Antihistamines are safe in pregnancy, but are likely to be effective in treating pruritus via their central sedative action. Histamine release has not been implicated in the mechanism of pruritus in these cases.

An ability to distinguish women with OC from those with more benign causes of pruritus in pregnancy based on the history of pruritus would be very useful. By definition OC has its onset in pregnancy, with no identifiable alternative cause and is rarely seen in the presence of a rash other than excoriations. On excluding women with a rash or alternative diagnosis, 57% remained ‘candidates’ for OC but only 2% of women had a final diagnosis of OC. The pruritus of OC is classically described as generalized and often worse on the palms of the hands and the soles of the feet,²⁴ suggesting that the location of pruritus may be discriminatory. Analysis of the completed questionnaires suggested selection of those women who report their pruritus to be either ‘all over’ or ‘most severe’ on the palms and/or soles would identify 48% of women with OC and 91% of disease-free women will be correctly reassured. Our group has confirmed this description in a larger cohort of 70 women with OC.²⁵ Eliciting such a history may be useful in the management of the pregnant woman with pruritus in identifying those women requiring more urgent investigation. We would therefore stress the need to always confirm or exclude the diagnosis biochemically. In view of recent studies in our group that have identified an abnormal lipid profile in women who later develop OC,²⁶ the distribution of pruritus in combination with a measurement of a specific biochemical index of OC such as dyslipidaemia may eventually prove useful in the clinic and lead to improved earlier diagnosis of OC prior to elevation of liver function tests and bile acids.

CONCLUSION

We have shown that pruritus in pregnancy is more common than previously reported and in the absence of a rash is often benign. OC occurred in 0.66% of the population studied. At present liver function tests should be performed in all pregnant women with pruritus in order to exclude this important cause of maternal and fetal morbidity and fetal mortality. The nature of the reported pruritus in a pregnant woman may however be helpful in identifying those women likely to have OC. Pruritus which is ‘all over’ or most severe on the palms and/or soles is particularly suggestive of the disorder.