Chapter 6. Bowel problems
Chapter Contents
Introduction121
Constipation121
Haemorrhoids123
Faecal incontinence125
Summary of the evidence used in this guideline133
What to do134
Summary guideline136
CONSTIPATION
Definition
Constipation is defined as difficulty in evacuating the bowel and is a common symptom during the second and third trimesters of pregnancy, arising from relaxation of the smooth muscle of the intestine as a result of circulating progesterone (Shepherd 2004). During the immediate postpartum period women may experience constipation following lack of dietary intake during labour or because of pain from perineal trauma.
Frequency of occurrence and risk factors
Although it is generally ‘well known’ that constipation can occur after childbirth, there is little documentary evidence of its prevalence or risk factors. Garcia & Marchant (1993), as part of an in-depth descriptive survey of postnatal care, sent 100 women a postal questionnaire at 8 weeks postpartum to ask about health problems; 90 women responded, 20 (22%) of whom had experienced constipation since their delivery, although onset and duration were not given. In a prospective observational study in Scotland, a representative sample of over 1200 women was surveyed whilst still in hospital, at 8 weeks and at 12–18 months, when half the sample was followed to ask about symptoms after 8 weeks: 19%, 20% and 7% respectively reported constipation at these various times (Glazener et al 1995). In a longitudinal survey by Saurel-Cubizolles et al (2000), 13% of the Italian women and 14% of the French reported constipation at 5 months. At 12 months these proportions were 17% and 26% respectively.
At all survey periods in the study by Glazener et al (1995), univariate analysis showed that constipation was significantly more common among women following instrumental deliveries, compared with spontaneous vaginal deliveries or caesarean sections. The proportions reporting constipation were, respectively, 31%, 17% and 20% in hospital, 31%, 18% and 17% between then and 8 weeks, and 14%, 6% and 5% after this. Constipation whilst still in hospital was found to be more common among primiparae (23%) than multiparae (16%), but after this the parity difference was non-significant (Glazener et al 1995). Since primiparae have much higher rates of instrumental deliveries than multiparae, the parity association may not be an independent one.
Schytt et al (2004) asked about constipation in a longitudinal study investigating urinary stress incontinence and found a prevalence of 21.7% and 28.6% in the third trimester among primiparae and multiparae respectively; these proportions were 25% and 29.3% at 4–8 weeks postpartum and 22% and 25.6% at 1 year. Constipation at 4–8 weeks was a significant independent predictor of stress incontinence at 1 year postpartum. French women who reported constipation at 12 months in the survey by Saurel-Cubizolles et al (2000) were more likely to be in employment, but there was no similar difference among the Italian women.
A longitudinal study in The Netherlands of 407 women, followed at 3 and 12 months postpartum to investigate defaecatory symptoms during and after a first pregnancy, found a much lower prevalence of constipation at 4.6% and 4.2% respectively but the definition (less than three bowel movements a week and need to strain 25% or more of the time) was more restrictive than in other studies. Independent risk factors for this were constipation at 12 weeks gestation and higher body mass index (BMI) (van Brummen et al 2006).
Management
There is limited research into the most effective treatment of constipation among either obstetric or non-obstetric populations and most conservative management of postpartum constipation is based on current clinical practice.
Jewell & Young (2001), in a Cochrane systematic review of interventions for the treatment of constipation during pregnancy, found two trials. One trial including 40 women showed that bran and wheat supplements were effective in increasing the frequency of defaecation (odds ratio (OR) 0.18, 95% confidence interval (CI) 0.05–0.67) and unlikely to have side-effects for the mother or fetus. The other trial, reporting on 140 women, found that stimulant laxatives were more effective than bulk-forming laxatives (OR 0.30, 95% CI 0.14–0.61) but may cause side-effects.
In an early randomised controlled trial in South Africa, comparing the use of the laxative senna (n = 224) with placebo (n = 247) for immediate postpartum constipation, senna was shown to be effective in producing spontaneous bowel action, but 12% of cases compared with 3% of controls experienced abdominal cramps, although most were described as mild (Shelton 1980).
Tramonte et al (1997) undertook a systematic review of general population trials (peripartum trials excluded) to evaluate whether laxatives and fibre therapy, for a minimum of 1 week, improved symptoms in adults who had experienced constipation for at least 2 weeks. Thirty six trials were included, with a total of 1815 subjects and a range of different laxative and fibre treatments. The reviewers concluded that both fibre and laxatives modestly improved bowel movement frequency but there was insufficient evidence to determine whether fibre was superior to laxative treatment or whether one class of laxative was better than another. They suggested that laxatives should only be given when simple treatments, such as fibre and dietary interventions, have failed.
HAEMORRHOIDS
Definition
Haemorrhoids result from swollen veins around the anus, which in severe cases can prolapse. Haemorrhoids can be graded according to the degree of prolapse: first-degree haemorrhoids are visible but do not prolapse; second-degree haemorrhoids prolapse with defaecation but return spontaneously; third-degree haemorrhoids prolapse and require manual replacement; and fourth-degree haemorrhoids remain prolapsed outside the anal canal (Pfenninger 1997). Haemorrhoids may occur during pregnancy as a result of the action of progesterone on the bowel, which increases varicosity, and during the postnatal period possibly as a consequence of pushing in the second stage of labour.
Frequency of occurrence and risk factors
Like constipation, haemorrhoids are considered to be a ‘well-known’ consequence of childbirth. Midwifery and obstetric textbooks have generally noted that haemorrhoids can cause a great deal of pain for a few days after delivery but then resolve, although they may worsen with subsequent pregnancies and can eventually become permanent (Hibbard 1988). More recently, however, observational studies of health problems after childbirth have found that haemorrhoids, whether of pregnancy or postpartum onset, do not always resolve quickly. MacArthur et al (1991), in an observational study, found that among 11,701 women 8% reported haemorrhoids of more than 6 weeks’ duration starting for the first time within 3 months of birth, and an additional 10% had ongoing or recurrent symptoms. Two-thirds of symptomatic women still had haemorrhoids at questioning, 1–9 years after delivery. Multivariate analysis showed that new symptoms were independently associated with forceps delivery, longer second-stage labour and vaginal delivery of a heavier baby. They were less likely after caesarean section and if delivered by section, there was no association with birthweight. These associations are all compatible with a longer or more expulsive period of pushing.
Glazener et al (1995), in a longitudinal study, found that 17% of women reported haemorrhoids (new and recurrent) when questioned in hospital, 22% between then and 8 weeks and 15% after this. As in the study by MacArthur et al (1991), haemorrhoids were significantly more common with an instrumental than a spontaneous vaginal delivery, and significantly less common with a caesarean section: 27%, 17% and 6% respectively whilst in hospital; and 31%, 23% and 14% between then and 8 weeks. At 12–18 months, however, no difference was found in haemorrhoids after 8 weeks in the spontaneous vaginal delivery (SVD) group (13%) relative to the caesarean section group (11%), but there were still significantly more in the instrumental delivery group (30%). Relationships with other factors, such as duration of second-stage labour or birthweight, were not reported.
Brown & Lumley (1998), in a population-based Australian survey at 6–7 months postpartum, found that haemorrhoids that women considered to be a problem for them were significantly associated with mode of delivery, occurring after 36% of instrumental deliveries, 25% of spontaneous vaginal deliveries, 17% of elective and 11% of emergency sections. Thompson et al (2002), in a longitudinal study in Australia, using similar questioning to Brown & Lumley (1998), found that the prevalence of haemorrhoids reduced over the first 6 months after birth from 30% between 0 and 8 weeks, to 18% at 16 weeks, but prevalence was still 13% between 16 and 24 weeks. Haemorrhoids were proportionately more common among women who had assisted compared with spontaneous vaginal deliveries and less common among those who had caesarean births, but these differences were not statistically significant. Ansara et al (2005), in a Canadian study, again using similar questioning to Brown & Lumley (1998), found the prevalence of haemorrhoids at 8–10 weeks postpartum to be 35.5%. After adjusting for several factors, including episiotomy and duration of labour, they found an independent positive association between haemorrhoids and instrumental relative to a spontaneous delivery (OR 2.57, 95% CI 1.16–5.70) and a negative association for caesarean section (OR 0.36, 95% CI 0.16–0.81).
In a European longitudinal study (Saurel-Cubizolles et al 2000) the prevalence of haemorrhoids among French and Italian women at 5 months postpartum was 16% in both cases, and prevalence at 12 months had increased to 21% among Italian and 26% among the French women. The authors suggest various reasons for this increase in symptoms, including a change in women’s perception of their health, so symptoms by then may feel more bothersome, as well as the effect of increasing demands of the baby.
Management
Treatment options available for haemorrhoids are varied and generally depend on severity, with scalpel surgery now reserved only for advanced fourth-degree cases (Pfenninger 1997).
Only one systematic review has been found specifically concerning postpartum populations. This review, published in 2005, is of ‘conservative management of symptomatic and/or complicated haemorrhoids in pregnancy and the puerperium’ (Quijano & Abalos 2005). Only two RCTs were found in childbearing populations, both of which were comparing phlebotonics (oral rutosides) with placebo during pregnancy, with no trials in postpartum populations. The authors concluded that, although this form of treatment looked promising for symptom relief in first- and second-degree haemorrhoids, its use cannot be recommended until new evidence provides reassurances about safety.
Like constipation, the conservative management of less severe childbirth-related haemorrhoids is based on current clinical practice. Relief of mild haemorrhoids may be provided by the topical application of creams, and dietary advice to ensure avoidance of constipation is also usually given, but there is no trial evidence of their effectiveness in either pregnant or postpartum populations. Alonso-Coello et al (2006) have recently completed a Cochrane review of laxatives for the treatment of haemorrhoids. This review found seven RCTs with a total of 378 participants, all in general populations, and concluded that laxatives in the form of fibre had a beneficial effect in the treatment of symptomatic haemorrhoids; the risk of not improving haemorrhoids and having persistent symptoms decreased by 53% in the fibre group (RR 0.47, 95% CI 0.32–0.68).
If haemorrhoids are severe and further treatment is required there are now numerous options, including rubber band ligation, cryothermy, injection sclerotherapy, infrared coagulation, diathermy and operative haemorrhoidectomy (MacRae and McLeod, 1995 and Pfenninger, 1997). No trials have specifically evaluated treatments for severe haemorrhoids in the postpartum period.
FAECAL INCONTINENCE
Definition
Faecal incontinence refers to the involuntary passage of bowel contents. The term is usually used to refer to solid or liquid faeces, which can be frank incontinence or soiling or staining of underwear, although faecal urgency is often also included. Faecal urgency is where the sensation of needing a bowel movement is felt but defaecation can only be deferred for a short period (a limit of 5 minutes or less is often given). Sometimes flatus incontinence is also included, although anal incontinence seems a more appropriate term used to refer to incontinence of faeces and/or flatus. There is currently no standard definition or classificatory system and no standard form of questioning, so differences in estimates of incidence and prevalence of faecal incontinence will occur. Where possible, the definition and method of ascertainment are described for each study.
Frequency of occurrence
It has long been recognised that childbirth is important in the pathogenesis of faecal incontinence based on histories of middle-aged women presenting in colorectal clinics, but until studies in the 1990s showed otherwise, it was generally assumed that women rarely (except after a third-degree tear) became symptomatic in the postpartum period (Swash 1993). In addition to epidemiological studies investigating prevalence and risk factors of faecal incontinence, there are data on symptom occurrence from pathophysiological studies of occult anal sphincter injury, and from studies primarily examining other problems (usually urinary incontinence). Since the last edition of this book, there has been a substantial increase in the number of studies on this topic.
Epidemiological studies
Sleep & Grant (1987), in a large randomised controlled trial of the effectiveness of pelvic floor exercises on urinary incontinence at 3 months postpartum, reported as an incidental finding that 3% of the women in the intervention and control groups experienced ‘occasional faecal loss’ but this was not mentioned further. Wilson et al (1996), in a study in New Zealand on the prevalence of urinary incontinence at 3 months postpartum, asked about faecal incontinence (new and recurrent symptoms, not including flatus) and found that 73 of the 1505 women (4.9%) experienced this, although they reported no further details.
An observational study in Birmingham, UK, of several health problems, including faecal incontinence, at 10 months postpartum obtained information on the incidence, duration and nature of this in a representative sample of 906 women (MacArthur et al 1997). Research midwives, in home-based interviews, asked about frank incontinence, soiling or staining of underwear, or urgency (described as ‘felt the need to go but couldn’t hold on’). Of the 906 women, 36 (4%) had experienced one or more of these as a new postpartum problem, the majority reporting onset as immediate or in the first 2 weeks. A further 19 women (2%) had recurrent or ongoing symptoms from a previous birth (n = 7) or because of a pre-existing condition (n = 12), mostly irritable bowel syndrome. Mean symptom duration among the 36 with new symptoms was 23 weeks; only five women had consulted their GP and 14 (39%) cases had resolved by the time of interview.
Zetterström et al (1999), in a prospective study in Sweden of anal incontinence among 278 primiparae who had a vaginal birth, found a low symptom frequency. At 5 months postpartum, five women (1.8%) reported incontinence of faeces and 70 (25%) of flatus, and at 9 months only three (1.1%) reported faecal and 71 (25%) flatus incontinence. A cross-sectional study from Israel (Groutz et al 1999) also found a low symptom frequency among 300 women interviewed at 3 months postpartum: 21 (7%) reported anal incontinence, two of faeces (0.7%), the remainder of flatus (6.3%). None had sought medical consultation. The European longitudinal survey of women’s health after childbirth found faecal incontinence rates at 5 months of 1% of 697 women in Italy and 3% of 589 women in France, and 3% and 5% respectively at 12 months (Saurel-Cubizolles et al 2000). Symptoms were self-reported with no definition of faecal incontinence given.
As part of a randomised controlled trial of treatment for postpartum urinary incontinence in high-risk women (instrumental births and baby 4000 g+) in Australia, data on faecal incontinence were also collected from the 568 women who were followed up at 12 months (Chiarelli et al 2003). Prevalence of faecal incontinence (ever accidentally lose solid or liquid stool) was 6.9%, 2.6% and 4.9% respectively, including some women with both. A multicentre study in England, Scotland and New Zealand also collected data on faecal incontinence from the whole study cohort (MacArthur et al 2001), whilst recruiting and following women as part of an RCT of treatment for urinary incontinence (Glazener et al 2001). At 3 months postpartum 7879 questionnaires were returned (72% response rate) and the prevalence of faecal incontinence (ever lose control of bowel motions from back passage in between visits to the toilet) was 9.6%. Six years after this birth the women were followed again, when 4214 responded. Of these, a total of 10% had faecal incontinence at that time; 3.6% had faecal incontinence on both occasions (MacArthur et al 2005). There was substantial resolution of symptoms and appearance of new ones: in 59% of the women who had symptoms at 3 months these had resolved, and 64% of women had symptoms at 6 years that had not been present at 3 months.
Several other recent studies have investigated postpartum faecal incontinence, with varying prevalence estimates. In a French study of 159 women who had instrumental deliveries and excluding women with previous anal incontinence, prevalence of loss of liquid or solid stool was 8.8% (Mazouni et al 2005). In a longitudinal study of 242 women up to 5 years following first delivery (excluding caesarean sections), prevalence of incontinence to solid or loose stool was 1% at 9 months and 5.4% at 5 years (Pollack et al 2004). A longitudinal study of 407 women followed at 3 and 12 months after birth in The Netherlands found prevalence of incontinence to liquid or solid stool to be 5.7% and 3.3% respectively (van Brummen et al 2006).
Pathophysiological studies
Investigative techniques developed in the 1990s, in particular endosonography and manometry, have enabled imaging of the anal sphincter muscles, and several studies of occult damage to internal or external anal sphincters have been undertaken. Most of these also obtained symptom data, but generally found higher symptom prevalences than in the epidemiological studies, probably because consent to the various anal investigations is more likely in symptomatic women. One study has demonstrated this (Chaliha et al 2001).
The first of these types of studies was undertaken by Sultan and colleagues (1993), who carried out examinations on 202 women who agreed to take part at 34 weeks’ gestation, and 150 who returned for further investigations at 6–8 weeks postpartum. Each time, the women completed a symptom questionnaire, which recorded anal incontinence, including faecal and flatus incontinence, and faecal urgency (defined as inability to defer defaecation for more than 5 minutes). Among the 150 women followed, 21 (14%) reported postpartum bowel control symptoms, 13 (8.7%) had new symptoms and eight (5.3%) recurrent or ongoing ones. Anal endosonography among the 79 primiparae who had delivered vaginally showed that none of these women had occult sphincter defects antenatally, but at 6–8 weeks 28 (35%) had defects. Among the multiparae 19 (40%) had defects antenatally and 22 (44%) at postnatal follow-up. The fact that postpartum incidence of defects among the primiparae was similar to the antenatal incidence of defects among multiparae indicates that anal sphincter damage is most likely following the first delivery. All except one of the women with anal incontinence or faecal urgency had a sphincter defect. Sphincter defects, however, were more common than symptoms; only about one-third of the women with defects experienced any loss of bowel control.
Similar studies followed that by Sultan et al and a meta-analysis to determine the incidence of obstetric anal sphincter damage and associated incidences of faecal incontinence was published in 2003 (Oberwalder et al 2003). It identified five studies published up to July 2001, including the one by Sultan et al (1993) described above. Based on a total of 717 vaginal deliveries, the incidence of anal sphincter defect in primiparae was found to be 26.9% and the incidence of new defects in multiparae was 8.5%. Overall, 29.7% of all women with defects were symptomatic (varying symptom definitions) and 3–4% of women experienced symptoms without a defect.
A large study, published since this meta-analysis, included 286 nulliparous women who underwent anorectal sensation and manometric examinations and completed a symptom questionnaire in the third trimester of pregnancy (Chaliha et al 2001). One hundred and sixty one women returned for the anorectal investigations and questionnaire completion at 3 months postpartum and the remaining 125 women completed the questionnaire by telephone. Overall prevalence of anal incontinence (including flatus) at 3 months was 10.5% and was 8.7%. for faecal urgency. Symptom prevalence was significantly greater in the women who returned for investigations.
Risk factors
In an early epidemiological study of 906 women by MacArthur et al (1997), obstetric data from case notes were obtained, and logistic regression showed that forceps and vacuum extraction delivery were the only maternal or obstetric factors that were independently associated with new faecal incontinence reported by the women. Zetterström et al (1999), in Sweden, found a relationship between instrumental delivery and anal incontinence at 5 months postpartum, but not at 9 months. This was a small study, however, and the authors noted that the 10% instrumental delivery rate among the primiparous sample was very low and may have affected these findings. Groutz et al (1999) also found a relationship between anal incontinence and instrumental delivery, although again the small sample size makes information on risk factors highly tentative. Moreover, almost all of the instrumental deliveries in both these studies were by vacuum extraction. A community-based general population study of over 6000 women (median age 58 years) showed that having ever had a forceps birth was significantly associated with faecal incontinence (OR 1.5), although how long after the birth the symptoms had begun was not available (Assassa et al 2000).
In the study of occult sphincter damage by Sultan et al (1993), the only obstetric factor independently associated with damage was forceps delivery: eight of the 10 women delivered by forceps had sphincter defects, although none of five delivered by vacuum extraction had defects. Donnelly et al (1998), in their Dublin study of sphincter damage in 124 primiparous women, found that instrumental delivery was associated with an 8.1-fold (95% CI 2.7–24.0) risk of anal sphincter injury and a 7.2-fold (95% CI 2.8–18.6) risk of symptoms (urgency, faecal and flatus incontinence). Abramowitz et al (2000), included in the earlier meta-analysis of anal sphincter damage, found forceps delivery to be an independent predictor of damage (OR 12, 95% CI 4–20).
Most studies are too small to allow a definite conclusion about the relative effects of instrumental delivery by forceps or by vacuum extraction. MacArthur et al (2001), in a large study of 4214 women at 3 months postpartum, were able to separately examine associations with each of the instruments. This showed that relative to spontaneous vaginal delivery, forceps delivery was independently associated with faecal incontinence (OR 1.94, 95% CI 1.29–1.30), but vacuum extraction was not (OR 1.26, 95% CI 0.77–2.07). A small trial randomised women who required an instrumental delivery to forceps (n = 61) or vacuum extraction (n = 69) and found that a significantly greater proportion of women in the forceps group reported altered faecal continence but no significant difference in continence score or faecal urgency was found (Fitzpatrick et al 2003). A small case–control study of 50 vacuum extraction and 50 spontaneous vaginal deliveries, matched for perineal trauma, showed that anal incontinence rates (urgency, flatus, liquid stool and soiling) were the same in both groups (Peschers et al 2003).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
