To describe the trends of severe perineal tears in England and to investigate to what extent the changes in related risk factors could explain the observed trends.
To describe the trends of severe perineal tears in England and to investigate to what extent the changes in related risk factors could explain the observed trends.
A retrospective cohort study of singleton deliveries from a national administrative database.
The English National Health Service between 1 April 2000 and 31 March 2012.
A cohort of 1 035 253 primiparous women who had a singleton, term, cephalic, vaginal birth.
Multivariable logistic regression was used to estimate the impact of financial year of birth (labelled by starting year), adjusting for major risk factors.
The rate of third-degree (anal sphincter is torn) or fourth-degree (anal sphincter as well as rectal mucosa are torn) perineal tears.
The rate of reported third- or fourth-degree perineal tears tripled from 1.8 to 5.9% during the study period. The rate of episiotomy varied between 30 and 36%. An increasing proportion of ventouse deliveries (from 67.8 to 78.6%) and non-instrumental deliveries (from 15.1 to 19.1%) were assisted by an episiotomy. A higher risk of third- or fourth-degree perineal tears was associated with a maternal age above 25 years, instrumental delivery (forceps and ventouse), especially without episiotomy, Asian ethnicity, a more affluent socio-economic status, higher birthweight, and shoulder dystocia.
Changes in major risk factors are unlikely explanations for the observed increase in the rate of third- or fourth-degree tears. The improved recognition of tears following the implementation of a standardised classification of perineal tears is the most likely explanation.
Recent population-based studies from Scandinavian countries and Canada have identified an increase in the occurrence of severe obstetric anal sphincter injuries.[1-5] In the UK, a study from a single unit reported that the combined rate of third-degree (anal sphincter is torn) and fourth-degree perineal tears (anal sphincter as well as rectal mucosa are torn) increased from 1.3% in 2001 to 4.6% in 2010. One possible reason for this trend is the rise in maternal age at first birth and maternal weight, which are linked to a higher birthweight and risk of perineal tears. Other reasons include increased awareness and training, which is likely to result in a better case detection and recording of obstetric injuries, and changes in the management of the second stage of labour.[1, 5]
The aim of this study was to describe the time trends in obstetric anal sphincter injuries in England, recorded in a large population-based database that includes all maternity admissions in the English National Health Service (NHS). We also investigated risk factors for these injuries and explored to what extent changes in these relevant risk factors and in obstetric practice were linked to the observed trends.
We used the Hospital Episode Statistics (HES) database to identify all deliveries that took place in English NHS Trusts (acute hospital organizations) from April 2000 to March 2012. HES is a data ‘warehouse’ that includes records of all inpatient admissions and day cases in English NHS Trusts. The data are extracted from local patient administration systems, and undergo a series of validation and cleaning processes before being made available for analysis.
The HES database contains patient demographics, clinical information, and administrative data for each inpatient episode of care. Diagnostic information is coded using the International Classification of Diseases 10th revision (ICD10), and operative procedures are coded using the UK Office for Population Censuses and Surveys classification, fourth revision (OPCS4). For maternity episodes, the HES database has supplementary fields known as the ‘maternity tail’, which captures parity, birthweight, gestational age, method of delivery, and pregnancy outcome. The accuracy and completeness of diagnostic and procedures data are high. The maternity tail is not compulsory, and the level of data completeness varies across Trusts. For example, birthweight and parity are available in 79 and 65% of the delivery episodes, respectively.
The study included only primiparous women aged 15–45 years, who had a singleton, term, cephalic, vaginal birth. We confined the analysis to NHS Trusts that had parity information recorded in at least half of the deliveries, and that had a proportion of primiparous women between 25 and 55% (overall about 40% of women giving birth are primiparous in England and Wales). The quality of parity data was evaluated for each year of the study.
Cases of perineal tears were identified by ICD10 codes O70.0 (first-degree perineal laceration), O70.1 (second degree), O70.2 (third degree), and O70.3 (fourth degree). Mode of delivery was defined using information in the OPCS4 procedure codes, and we distinguished between vaginal (OPCS4 codes R23 and R24), forceps (R21), and ventouse (R22), or if not defined using OPCS4 codes, by the delivery method specified in the maternity tail. These three modes were further stratified by whether or not an episiotomy had been performed (OPCS4 code R27.1).
We identified the following potential risk factors. Maternal demographic factors were age (<20, 20–24, 25–29, 30–34, ≥35 years), ethnicity (white, Asian, black, other), and socio-economic deprivation of the mother's area of residence using the index of multiple deprivation (IMD, quintiles of 32 480 areas in England ranked according to a measure of deprivation that combines a range of economic, social, and housing indicators). Factors associated with labour were mode of delivery, birthweight, prolonged labour, and shoulder dystocia. The duration of labour was marked as prolonged if the delivery record included an ICD10 diagnosis code O63 (long labour), whereas shoulder dystocia was identified by the ICD10 code O66.0 (obstructed labour as a result of shoulder dystocia).
We present the unadjusted trends for all degrees of perineal tears for all singleton, term, cephalic, vaginal first births. We used multiple logistic regression models to estimate the crude and adjusted effect of the financial year of delivery (labelled by starting year) on the risk of observing a third- or fourth-degree tear, with the aim of assessing to what extent the magnitude of the calendar time effect is mitigated by controlling for the other risk factors. The logistic regression model was defined with random intercepts at the NHS Trust level to take account of organisational clustering. All analyses were performed in Stata/SE 11.
There were 6 621 439 singleton term deliveries in 146 English NHS Trusts between April 2000 and March 2012. Among these, 39.1% took place in NHS Trusts that had poor-quality parity data, and the records for these NHS Trusts were omitted. The median number of NHS trusts included in each year was 81 (interquartile range: 79–85). Omitting episodes with missing parity data left 3 559 687 deliveries, of which 1 358 072 (38.6%) were first births. Among these primiparous women, 23.1% of deliveries were by caesarean section, and 0.2% were vaginal breech deliveries. A further 0.6% of records were missing maternal age or deprivation data. Excluding these left 1 035 253 deliveries for analysis.
The trends in unadjusted rates of reported obstetric tears at first births, by degree of tear, are given in Figure 1. The rate of third- or fourth-degree tears tripled between 2000 and 2011, whereas the rate of second-degree tears increased by 23.5%. In 2011, the rate of third- or fourth-degree tears was 5.9 per 100 deliveries.
During the same period, the use of forceps among all vaginal primiparous deliveries increased from 9.0 to 16.1%, and the rate of ventouse deliveries decreased from 17.5 to 13.9% (Figure 2). Only 83.2% of forceps deliveries were facilitated by episiotomy, with the rate increasing from 82.2 to 87.7% over the study period. The proportion of ventouse deliveries facilitated by episiotomy increased from 67.8% in 2000 to 78.6% in 2011. The use of episiotomy in non-instrumental deliveries decreased over the study period from 19.1 to 15.1%.
Over half of the women included in the study were between 20 and 29 years of age (Table 1). The risk of a third- or fourth-degree tear increased with maternal age. Women older than 25 years were reported to have a third- or fourth-degree tear at least twice as often as teenage mothers. Women living in the least deprived communities, and those with non-white ethnicities were also more likely to have a severe obstetric tear. Asian women had a risk of a third- or fourth-degree tear that was more than twice as high as women from a white ethnic background (adjusted OR 2.27, 95% CI 2.14–2.41).
|Prevalence of risk factor (%)||Rate of tear per 100 births (%)||Crude OR (95% CI)||Adjusted OR (95% CI)|
|Year of delivery (Financial years)|
|2001||2.0||1.16 (1.02–1.31)||1.13 (0.99–1.29)|
|2002||2.4||1.35 (1.19–1.52)||1.32 (1.17–1.49)|
|2003||2.8||1.63 (1.43–1.85)||1.57 (1.37–1.80)|
|2004||3.0||1.75 (1.56–1.98)||1.68 (1.48–1.90)|
|2005||3.7||2.14 (1.91–2.39)||2.02 (1.79–2.26)|
|2006||4.2||2.48 (2.19–2.81)||2.29 (2.01–2.62)|
|2007||4.7||2.74 (2.41–3.12)||2.48 (2.14–2.87)|
|2008||4.9||2.85 (2.52–3.23)||2.56 (2.21–2.96)|
|2009||5.1||2.98 (2.61–3.41)||2.70 (2.33–3.13)|
|2010||5.6||3.31 (2.90–3.77)||3.02 (2.63–3.45)|
|2011||5.9||3.48 (3.03–3.99)||3.15 (2.74–3.62)|
|Maternal age (years)|
|<20||15.3||2.0||0.40 (0.38–0.42)||0.50 (0.48–0.53)|
|20–24||26.4||3.2||0.66 (0.63–0.68)||0.70 (0.68–0.73)|
|30–34||22.1||5.0||1.07 (1.03–1.10)||1.07 (1.04–1.11)|
|≥35||8.3||4.5||0.95 (0.89–1.01)||0.93 (0.89–0.98)|
|Asian||9.2||6.1||1.66 (1.50–1.84)||2.27 (2.14–2.41)|
|Black||3.2||3.9||1.04 (0.94–1.15)||1.32 (1.22–1.44)|
|Other||4.5||4.4||1.19 (1.10–1.29)||1.26 (1.17–1.36)|
|Missing||10.9||3.3||0.88 (0.82–0.95)||0.98 (0.93–1.05)|
|Q1: Most deprived||26.9||3.4||–||–|
|Q2||21.8||3.7||1.11 (1.05–1.17)||1.01 (0.97–1.05)|
|Q3||18.5||4.1||1.24 (1.17–1.32)||1.06 (1.01–1.12)|
|Q4||16.6||4.2||1.26 (1.18–1.35)||1.06 (1.01–1.12)|
|Least deprived||16.2||4.6||1.39 (1.27–1.53)||1.14 (1.07–1.21)|
|Mode of delivery|
|Normal w/o episiotomy||61.0||3.4||–||–|
|Normal w/episiotomy||11.3||2.2||0.63 (0.58–0.69)||0.57 (0.51–0.63)|
|Forceps w/o episiotomy||1.9||22.7||8.30 (7.10–9.70)||6.53 (5.57–7.64)|
|Forceps w/episiotomy||9.8||6.1||1.84 (1.67–2.01)||1.34 (1.21–1.49)|
|Ventouse w/o episiotomy||4.7||6.4||1.94 (1.79–2.10)||1.89 (1.74–2.05)|
|Ventouse w/episiotomy||11.2||2.3||0.67 (0.61–0.74)||0.57 (0.51–0.63)|
|<2500||3.3||1.4||0.38 (0.34–0.42)||0.37 (0.33–0.40)|
|2500–3000||18.3||2.4||0.68 (0.65–0.71)||0.66 (0.64–0.69)|
|3501–4000||26.6||5.1||1.49 (1.45–1.54)||1.50 (1.46–1.54)|
|>4000||7.0||7.8||2.36 (2.26–2.47)||2.27 (2.18–2.36)|
|Missing||5.3||3.6||1.03 (0.84–1.26)||1.15 (0.94–1.40)|
|Yes||16.7||5.4||1.49 (1.4–1.59)||0.99 (0.94–1.04)|
|Yes||0.8||11.3||3.15 (2.9–3.43)||1.90 (1.72–2.08)|
Women who had an episiotomy were less likely to experience a severe perineal tear, regardless of the mode of delivery. Across the different modes of delivery, women who had a non-instrumental or a ventouse delivery with an episiotomy had the lowest rates of third- or fourth-degree tears. Use of forceps increased the risk of a tear, with a forceps delivery without an episiotomy increasing the odds of a tear six-fold compared with a vaginal delivery without an episiotomy. The adjusted risk of third- or fourth-degree tears increased with birthweight and shoulder dystocia, but was not associated with the duration of labour.
Figure 3 shows the time trends within risk groups according to maternal age, ethnicity, mode of delivery, and shoulder dystocia. The rate of obstetric tears increased in all groups, with the largest absolute increase in women undergoing a forceps delivery without an episiotomy (Figure 3c) and in women with an Asian ethnic background (Figure 3b).
We found a three-fold increase in the rate of reported third- or fourth-degree perineal tears in England, with the rate rising from 1.8% in 2000 to 5.9% in 2011. An increased risk of a severe tear was associated with a maternal age above 25 years, forceps and ventouse delivery, especially without episiotomy, Asian ethnicity, a more affluent socio-economic status, higher birthweight, and shoulder dystocia. The use of an episiotomy was protective; however, the increase in the rate of severe perineal injury over the study period could not be explained by temporal changes in the major risk factors.
Using HES data has several advantages for trying to describe patterns of maternity care. First, over 96% of all deliveries in England occur in NHS Trusts, and are therefore captured by HES, which gives large sample sizes for outcomes that are relatively rare, such as third- or fourth-degree perineal tears. Second, the availability of data since 1997 allows for the analysis of patterns of care over time. Finally, the data are able to capture multiple procedures and diagnoses at an individual level, and so provide a rich description of the patient case mix.
A weakness of administrative data sets is that the coding of the diagnoses and procedures is potentially inaccurate; however, studies have demonstrated that the majority of NHS Trusts submit good-quality data to HES that conforms with national recommendations.[13-15] A recent systematic review of discharge coding accuracy in the UK concluded that routinely collected data are sufficiently robust to support their use for research and managerial decision-making. The richness of the data also makes it possible to develop coding frameworks and data quality criteria to identify hospitals with divergent coding practices by combining diagnosis, procedure, and administrative codes. A number of recent publications have demonstrated that when analysed carefully, HES is a valuable source of data to explore patterns of care as well as supporting epidemiological studies related to childbirth.[17-19]
This study included half of all vaginal singleton term births in primiparous women who delivered in an NHS hospital over a 12–year period. We focused on primiparous women, as earlier studies had concluded that birth order and a perineal tear in an earlier birth are important risk factors.[20-26] Excluding NHS Trusts with poor data quality may have introduced bias, as the risk-adjusted tear rates at these hospitals may be different from the rates observed in hospitals with better data quality. However, the effect size of selection bias is likely to be small because the distributions of outcome and risk factors in both groups were similar (Table S1). This finding is in agreement with a recent study that concluded that using birth cohorts from hospitals with high completeness of recording is likely to be valid and nationally representative. A second method to identify parity is to examine the women's obstetric history. In a sensitivity analysis, we constructed a data set of primiparous women, using 10 years of obstetric history, covering a study period from April 2007 to March 2012. The adjusted estimates of risk of obstetric tears from this data set were compared with the results from the data set generated using high-quality parity information for the same period. Both methods yielded comparable results (data not shown).
We were unable to control for a number of risk factors, such as the type of anaesthetic used, that might have influenced our results. Data on intrapartum anaesthetic use is available in HES, but this information is contained in the maternity tail, and was missing in about one-third of all patient records. Therefore, we omitted this variable from the analysis. In the subsample of records for whom this information was available, the adjusted effect of epidural analgesia was OR 1.10 (95% CI 0.94–1.29), and the inclusion of epidural in the logistic regression model did not significantly modify the effect size of other risk factors. This result is consistent with other studies on the impact of an epidural anaesthetic on third- or fourth-degree tears.[28, 29] We were also unable to control for perineal protection techniques applied during the second stage of labour, or the experience or preferences of the birth attendant.[30, 31] Similarly, the angle and size of an episiotomy is likely to influence the risk of tears,[32-34] but this information was not available in our data set.
The rate of reported third- or fourth-degree tears in singleton, term, cephalic, vaginal first births in England was 5.9% in 2011. This rate of clinically recognised anal sphincter injuries falls within the wide range of figures reported elsewhere. In large population-based studies using birth registry or administrative hospital data, the incidence was 1.8% in Finland, 3.6–4.2% in Norway, Denmark, and Sweden,[1, 2] and 4.5–5.4% in the USA.[20, 35] It is known that the actual rate of anal sphincter lacerations is significantly higher than the reported rates. Studies using endoanal ultrasonography have found clinically occult anal sphincter defects in up to one-third of vaginal deliveries.
We found that the risk of perineal tears was lower in younger women. The risk of a severe perineal injury in teenage primiparous women was less than half the risk in women older than 25 years, which corresponds to the rates in Norway. However, maternal age was not always identified as a risk factor for severe tears.[28, 37, 38]
Differences in risk for ethnicity have been demonstrated in studies from Norway, Sweden, UK, and the USA.[1, 2, 20, 38-40] It has been suggested that differences in the anatomy of the perineum, such as perineal body length and thickness among different ethnic groups, may be contributing factors.
Mode of delivery is a key determinant of the risk of perineal tears, with studies consistently demonstrating that women with instrumental deliveries have higher rates of anal sphincter tears,[2, 23, 28, 37] and that forceps deliveries carry the highest risk of third- or fourth-degree perineal tears. The risk of having a severe perineal injury has been reported to be 1.5–14.0 times higher with forceps, and up to four times higher with ventouse, than with spontaneous vaginal delivery.[21-23, 38, 41, 42]
We considered it more informative to analyse combinations of mode of delivery and use of episiotomy in contrast to analysing both as separate risk factors, which has been the case in most studies. This allows for the effect of episiotomy to vary by delivery mode. Midline episiotomies are known to increase the risk of third- or fourth-degree perineal tears.[21, 22, 41, 42] For mediolateral episiotomies, although the evidence is not conclusive,[43, 44] most studies suggest that this technique protects against severe tears.[20, 23, 28, 29, 32, 37, 41, 45-47] The results of studies that analysed specific combinations of mode of delivery and episiotomy use were consistent with ours. These studies found that mediolateral episiotomy reduced the risk of tears in instrumental vaginal deliveries.[1, 46, 47]
Our findings on shoulder dystocia and birthweight confirms the results of previous studies, which found that shoulder dystocia and birthweights higher than 4000 g double the risk of perineal tears.[20-23, 45, 48] An increase in the incidence of these risk factors could contribute to a higher rate of tears. However, the distribution of birthweights in our population did not change over the study period. In fact, the use of episiotomy in instrumental deliveries for babies with birthweights over 4000 g increased from 77.5% in 2000 to 85.6% in 2011, which is likely to reduce the risk of severe tears for this group. We did not find evidence that a longer duration of labour increases the risk of severe tears, which is in contrast to a number of other studies.[23, 37, 45, 48]
It is important to monitor trends in the incidence of third- or fourth-degree perineal tears, and the underlying explanations, because severe perineal trauma is listed as an index of quality of care in the RCOG Maternity Dashboard, and by Australian, European, and US national quality accreditation systems. These nationally reported trends can be used for benchmarking. A trend towards an increasing incidence of third- or fourth-degree perineal tears, as found in this study, does not necessarily indicate poor-quality care. It may indicate, at least in the short term, an improved quality of care through better detection and reporting.
The most likely explanation for the rising rate of reported severe perineal injury is improved recognition. This would be a result of two recent developments: the introduction of a standardised classification of perineal tears, and better training of staff in recognising and repairing perineal tears. The Royal College of Obstetricians and Gynaecologists published evidence-based guidelines for the management of third- or fourth-degree perineal tears in 2001 (second edition in 2007). All maternity units in England should now have written policies on the diagnosis and management of tears. Prior to the introduction of the standardised classification, some clinicians will have classified injuries to the anal sphincter as second-degree tears.[31, 55, 57] In the last decade, specific training in the identification and repair of perineal tears has become established as an essential component of postgraduate training and continuing professional development for doctors and midwives. Studies in the UK that have evaluated the implementation of the documentation proforma and auditable standards recommended in the new guideline,[58, 59] and in training interventions, confirm that the increased awareness and appropriate examination have increased the likelihood of perineal tears being detected.
Another possible explanation is a gradual improvement in the coding of tears in the English HES database. However, better coding is unlikely to have had a major impact as the completeness and accuracy of data coding of third- or fourth-degree perineal tears were found to be high in databases in the USA, Norway, and Australia.[62-64] In all these countries the sensitivity of coding of third- or fourth-degree tears was higher than 90%, and the majority of discrepancies occurred in the coding of first- and second-degree tears.
Our results – and those of other studies – demonstrate that changes in the main risk factors do not explain the observed increase in the rates of severe perineal tears.[1-3, 5, 35] However, there have been significant changes in the management of the second stage of labour in the last decade. In the 1990s, ventouse was advocated as ‘the instrument of first choice’ for instrumental vaginal delivery.[65, 66] As the rate of failed instrumental delivery increased, clinical guidelines moved to recommending the use of the instrument best suited to the individual circumstances. The National Institute for Clinical Excellence (NICE) Guidelines for intrapartum care also recommended that routine episiotomy should not be performed during spontaneous vaginal birth, but that it should be used with any forceps delivery. These changes, as well as the fact that an episiotomy was not performed in one or two of every ten forceps deliveries in our study population, may have contributed to the increase in the rates of third- or fourth-degree tears in England.
Changes in the application of perineal protection techniques may also have played a role.[68-71] The implementation of manual assistance and perineal protection techniques during the second stage of labour have significantly reduced the incidence of perineal tears in Norway.[72, 73] Antenatal perineal massage reduces the likelihood of perineal trauma (mainly episiotomies), but is not routinely practiced in the UK. Wider application of the ‘hands-poised’ approach, combined with the reluctance to use episiotomies, could have resulted in a higher risk of a third- or fourth-degree tears.[3, 75, 76] Also, women are increasingly encouraged to use their preferred birth positions, which may have reduced perineal protection.[5, 48]
This study found that, between April 2000 and March 2012, the rate of reported third- or fourth-degree perineal tears for first births tripled in England. This trend mirrors those reported from other developed countries such as Finland, Norway, and Canada. The most likely explanation for the increasing rate is improved diagnosis through the introduction of a standardised classification of perineal tears and the better training of staff. Changes in the patterns of maternal risk factors and modes of delivery are unlikely explanations.
IGU, LCE, TAM, LA, and JHvdM conceived the study. IGU and DAC contributed to its design and conducted the analyses. IGU wrote the article, and DAC, LCE, TAM, LA, DR, AT, and JHvdM commented on drafts. All authors approved the final version for publication.
The study is exempt from UK National Research Ethics Service approval because it involved the analysis of an existing data set of anonymised data for service evaluation. Approvals for the use of HES data were obtained as part of the standard Hospitals Episode Statistics approval process.
IG-U is supported by the Royal College of Obstetricians and Gynaecologists.
We thank the Department of Health for providing the HES data used in this study.