Int Urogynecol J DOI 10.1007/s00192-014-2478-7

ORIGINAL ARTICLE

Risk factors for obstetric anal sphincter injuries and postpartum anal and urinary incontinence: a case–control trial Madeline Burrell & Sapna Dilgir & Vicki Patton & Katrina Parkin & Emmanuel Karantanis

Received: 5 April 2014 / Accepted: 29 June 2014 # The International Urogynecological Association 2014

Keywords Anal incontinence . OASIS . Obstetric anal sphincter injury . Risk factors

least 1.7–3 % of deliveries [1, 2]. OASIS involves the disruption of the anal sphincter and musculature and/or the anal mucosa during vaginal birth. The severity of the injury can be graded using the Sultan classification for 3rd and 4th degree tears [3, 4]. OASIS has been linked with various complications such as anal incontinence, urinary incontinence, perineal pain, dyspareunia, embarrassment and low self-esteem [5, 6]. Childbirth is the main cause of anal incontinence amongst women, who are nine times more likely to experience incontinence than men [7, 8]. Postpartum urinary incontinence occurs in 7–34 % of women, but is especially present in women sustaining OASIS, 13– 46 % of whom suffer post-partum urinary incontinence [9]. Thus, it would be far more beneficial to the patient if OASIS could be prevented. By identifying the risk factors during childbirth, doctors and midwives may be able to recognise women at risk and investigate methods to prevent the occurrence of OASIS. This study aimed to analyse a wide range of potential factors for OASIS, including potential racial origin and birth factors, by collecting birth data as soon as possible after the birth from the accoucheur, to attempt to obtain finer detail of the birth factors compared with typical factors collected in a hospital database.

Introduction

Materials and methods

Although perineal trauma is common during childbirth, obstetric anal sphincter injury (OASIS) is not, accounting for at

Recruitment of OASIS and control groups

Abstract Introduction and hypothesis Obstetric anal sphincter injuries (OASIS) cause serious maternal morbidity for mothers. A clearer understanding of aetiological factors is needed. We aimed to determine the risk factors for OASIS . Methods Birth details of 222 primiparous women sustaining OASIS were compared with 174 women who did not sustain OASIS (controls) to determine the relevant risk factors. The data underwent univariate analysis and logistic regression analysis. Results Asian or Indian ethnicity, operative vaginal birth (p=0.00), persistent occipito-posterior position (p=0.038) and rapid uncontrolled delivery of the head were identified as risk factors for OASIS. Pushing time, use of epidural, episiotomy and head circumference were not predictors of OASIS. Conclusions Women with Asian or Indian ethnicity, operative vaginal birth, persistent occipito-posterior position and rapid uncontrolled delivery of the fetal head were likely to sustain OASIS. Awareness of these factors may help to minimise the incidence of OASIS.

M. Burrell : S. Dilgir : V. Patton : K. Parkin : E. Karantanis (*) Pelvic Floor Unit, St. George Hospital, University of New South Wales, Sydney 2217, Australia e-mail: [email protected]

Women in this study were recruited from a tertiary university hospital in Sydney, Australia, serving a diverse multicultural population in a public setting. Two groups were studied in this case–control study: a group who sustained OASIS (cases) and a group who did not (controls). Women from non-English

Int Urogynecol J

speaking backgrounds were not excluded and the national interpreter service was used when required. Local ethics approval was obtained for this study (HREC/10/STGH/81).

Table 1 Comparison of demographic characteristics of the study sample with obstetric anal sphincter injuries (OASIS) with the women with OASIS in the hospital database (2011–2012)

OASIS group Existing data from women who sustained OASIS in the hospital and who were seen in the hospital’s perineal clinic were used for the study. OASIS was identified by direct visualisation and confirmed by rectal examination if in doubt, either by the accoucheur, or by a registrar. Ultrasound was not used to assist diagnosis. Sultan’s classification of perineal tears [4] was used to assess the severity of the OASIS group’s injuries, as determined by those witnessing the tears and documenting on a standardised proforma. This proforma is completed for every woman sustaining OASIS as part of the routine hospital protocol that has been in place for all women sustaining OASIS since 2006. Such data were collected by the accoucheur, within an hour of the birth or correction of the injury. The variables recorded included race, use of operative vaginal delivery (vacuum or forceps), the use of episiotomy, duration of pushing, use of epidural and head position at birth such as occipito-anterior (OA) or occipito-posterior (OP) or lateral, and treatment details. Although a persistent OP position at birth (face to pubis) or other positions at the delivery of the head were recorded, any attempt at manual rotation before delivery was not documented. Rotational forceps were not used in our institution during this time. We defined our racial groups as: Asian (Chinese, South-East Asian); Indian (from India, Bangladesh, Nepal and Sri Lanka); Caucasian (European; Middle Eastern); African; and Polynesian. Women with OASIS were managed according to hospital protocol and were followed up and investigated in a dedicated perineal clinic. The proforma that was completed at birth was viewed in the perineal clinic and data were entered into a dedicated database. Symptom severity at 6 weeks postpartum was assessed using fully psychometrically tested instruments. Anal incontinence severity was recorded using a St Marks’ Continence Score [10]. Urinary incontinence severity was recorded using the ICIQ-SF [11]. Birth data from women who sustained OASIS and who did not attend the clinic were not included in this study. The degree to which the demographic details of our OASIS study group were representative of women with OASIS in the hospital database in 2011–2012 was studied (Table 1). Control group The control group was recruited from the delivery suite between 2010 and 2013 (primarily by authors MB and SD) and birth data were collected using the same proforma as for the OASIS group. An ethics committee-approved consent form

Age, years (±SD) Birth weight, g (±SD) Term, 37–41 weeks (%) Asian race (%) Indian race (%) Operative delivery (%)

OASIS sample (n=222)

Hospital sample (n=163)

P

29.4 (4.6) 3,405 (497) 177 (81.6) 93 (41.9) 31 (13.9) 108 (48.4)

28.3 (3.8) 3,393 (410) 128 (78.53) 61 (37.4) 46 (28.2) 94 (57.7)

0.0131 0.797 0.840 0.407 0.0007 0.263

was developed for women who did not sustain OASIS. Nulliparous women in early labour were approached to consent to their labour information being used, and to being called at 6 weeks post-partum for a symptom assessment using the same proforma and outcome measures as those used for women sustaining OASIS. Consenting women were recruited in the delivery suite while in labour, or they consented soon after the birth if they were not in a position to consent beforehand. Although a few women refused, the number of those refusing was not documented. Women in the control group were approached sequentially on set days of the week with the intention of including all consenting women on that particular day, even if they were non-English-speaking, in the hope of an accurate representation of the population of women who did not sustain OASIS. Only women with a confirmed OASIS or who had an emergency caesarean, or who did not consent were excluded. The importance of not missing OASIS injuries and the use of rectal examinations to confirm a diagnosis have been emphasised to hospital staff over the last 4 years with a resultant increase in the OASIS detection rate since 2008. Although occult sphincter injuries are possible, they are rare and can usually be detected at birth [12], but no other attempt was made to detect such injuries. In particular ultrasound was not used in the acute phase to detect occult OASIS. The degree to which our control group was representative of hospital patients was determined by comparing demographic and birth characteristics of our controls with those of women without OASIS from the hospital’s obstetric database for the years 2011–2012 (Table 2). Body mass index is a notable omission, as this was not tested in controls and could not be compared. While data from our controls described racial origin, hospital data gave only information on the country of origin. Thus, in the hospital database assumptions were made about racial origin. For example, if a woman was born in China, we assumed she was Asian. This may have underestimated the number of Asian and Indian women at hospital, as 655 (39.9 %) women were “born in Australia,” a proportion of whom would be Asian, Indian, etc.

Int Urogynecol J Table 2 Comparison of the control group with primiparous women delivering vaginally without OASIS in the hospital database (2011–2012)

Mean age (±SD) Term, 37–41weeks (%) Mean birth weight, g (±SD) Asian (%) Indian (%) Operative delivery (%) 1st/2nd degree tear (%)

Controls (n=174)

Hospital population P value (n=1,640)

29 (4.7) 110 (82.7) 3,277 (455) 30 (17.2) 25 (14.4) 36 (20.6) 75 (43.1)

28 (4.8) 1,482 (90) 3,222 (518) 440 (22.9) 249 (12.9) 523 (31.8) 734 (44.8)

0.19 0.61 0.22 0.006 0.99 0.285 0.689

At 6 weeks post-partum, the control group was contacted by telephone or emailed a survey enquiring about urinary incontinence, anal incontinence and sexual activity. The St Mark’s Continence score and ICIQ-SF score were again used to investigate incontinence, but data from these findings are the subject of a future publication. Statistical analysis Data were entered into SPSS (version 22) and potential birth factors underwent univariate analysis with logistic regression analysis being undertaken for those factors that appeared to be independent variables. Continuous variables were analysed using independent samples t test with significance designated at 0.05. Categorical variables were analysed using Chisquared tests and odds ratios were calculated. Stratification of analysis was performed to observe the effects of interventions that might modify the incidence of OASIS, such as instrumental delivery and episiotomy. For each variable, missing data were not included in the calculations. Power analysis was performed in our study to determine sample size using hospital statistics from 2011 to 2012 (obtained from the ObstetriX database) of primiparous women delivering vaginally. Based on these data, for a prevalence of Asian-only women of 37.4 % with OASIS and 22.9 % without we required 157 cases of OASIS and 157 controls. For Indianonly women with a hospital prevalence of OASIS of 28.2 % and non-OASIS prevalence of 12.9 %, we required 122 cases and 122 controls.

Results One hundred and seventy-four women who gave birth to their first child by vaginal birth were recruited into the control group. Racial and birth data were analysed using information from the births of all 174 women in the control group. Postpartum data were analysed using the sample of 133 women

who agreed to follow-up (but are to be presented in a future manuscript). Data from 222 women (mean age =29.42) who sustained OASIS and attended the perineal clinic were available for analysis. The proportion of women sustaining 3a tears was 46.2 %, 3b 25.8 %, 3c 14 %, and 14 representing unclassified 3rd degree or 4th degree tears. Validation of study groups The hospital database used to validate the study groups was a sample of 1,803 primiparous women who delivered vaginally in 2011–2012 (not including vaginal birth after caesarean). One hundred and sixty-three of these women (8.9 %) had sustained a documented OASIS, while 1,640 had not. General birth events data of women with OASIS in the study group and the hospital database (Table 1) showed that data from our study sample were similar to the representative hospital population. There was a statistically significant but clinically insignificant difference in age between groups that we feel reflected the lag between age during childbirth and age at postnatal review. Racial data suggested that the proportion of Asian women might be similar in both groups. Importantly, however, the proportion of Indian women in the study sample was significantly lower than the hospital sample, suggesting that these women might have had a tendency not to attend the perineal clinic for review at their 6-week postnatal visit. The comparison of general birth events between our control group and hospital primiparous women without OASIS found no significant differences (Table 2). However, a significantly smaller proportion of the recruited control group were Asian compared with the hospital database sample. This would suggest that Asian women might either have been avoided during recruitment, or had been more likely to refuse consent. This limitation of the study may exaggerate the difference in Asian prevalence between our study and control groups. Therefore, for information on racial prevalence, data from the hospital database and not the study groups was relied upon, indicating that, of women in hospital sustaining OASIS, 28.2 % were of Indian racial origin, whereas only 12.9 % of women who did not sustain OASIS were Indian (P60 min before a spontaneous vaginal birth were more prevalent in the OASIS group than those pushing for

Risk factors for obstetric anal sphincter injuries and postpartum anal and urinary incontinence: a case-control trial.

Obstetric anal sphincter injuries (OASIS) cause serious maternal morbidity for mothers. A clearer understanding of aetiological factors is needed. We ...
181KB Sizes 0 Downloads 5 Views