Risk of recurrence and subsequent delivery after obstetric anal sphincter injuries

Authors

  • E Baghestan,

    1. Institute of Clinical Medicine, University of Bergen, Bergen, Norway
    2. Department of Obstetrics and Gynaecology, Haukeland University Hospital, Bergen, Norway
    3. Medical Birth Registry of Norway, Norwegian Institute of Public Health, Bergen, Norway
    Search for more papers by this author
  • LM Irgens,

    1. Medical Birth Registry of Norway, Norwegian Institute of Public Health, Bergen, Norway
    2. Locus for Registry Based Epidemiology, Department of Public Health and Primary Health Care, University of Bergen, Norway
    Search for more papers by this author
  • PE Børdahl,

    1. Institute of Clinical Medicine, University of Bergen, Bergen, Norway
    2. Department of Obstetrics and Gynaecology, Haukeland University Hospital, Bergen, Norway
    Search for more papers by this author
  • S Rasmussen

    1. Institute of Clinical Medicine, University of Bergen, Bergen, Norway
    2. Department of Obstetrics and Gynaecology, Haukeland University Hospital, Bergen, Norway
    3. Medical Birth Registry of Norway, Norwegian Institute of Public Health, Bergen, Norway
    Search for more papers by this author

Dr E Baghestan, Department of Obstetrics and Gynaecology, Haukeland University Hospital, N-5021, Bergen, Norway. Email elham.baghestan@kk.uib.no

Abstract

Please cite this paper as: Baghestan E, Irgens L, Børdahl P, Rasmussen S. Risk of recurrence and subsequent delivery after obstetric anal sphincter injuries. BJOG 2012;119:62–69.

Objective  To investigate the recurrence risk, the likelihood of having further deliveries and mode of delivery after third to fourth degree obstetric anal sphincter injuries (OASIS).

Design  Population-based cohort study.

Setting  The Medical Birth Registry of Norway.

Population  A cohort of 828 864 mothers with singleton, vertex-presenting infants, weighing 500 g or more, during the period 1967–2004.

Methods  Comparison of women with and without a history of OASIS with respect to the occurrence of OASIS, subsequent delivery rate and planned caesarean rate.

Main outcome measures  OASIS in second and third deliveries, subsequent delivery rate and mode of delivery.

Results  Adjusted odds ratios of the recurrence of OASIS in women with a history of OASIS in the first, and in both the first and second deliveries, were 4.2 (95% CI 3.9–4.5; 5.6%) and 10.6 (95% CI 6.2–18.1; 9.5%), respectively, relative to women without a history of OASIS. Instrumental deliveries, in particular forceps deliveries, birthweights of 3500 g or more and large maternity units were associated with a recurrence of OASIS. Instrumental delivery did not further increase the excess recurrence risk associated with high birthweight. A man who fathered a child whose delivery was complicated by OASIS was more likely to father another child whose delivery was complicated by OASIS in another woman who gave birth in the same maternity unit (adjusted OR 2.1; 95% CI 1.2–3.7; 5.6%). However, if the deliveries took place in different maternity units, the recurrence risk was not significantly increased (OR 1.3; 95% CI 0.8–2.1; 4.4%). The subsequent delivery rate was not different in women with and without previous OASIS, whereas women with a previous OASIS were more often scheduled to caesarean delivery.

Conclusion  Recurrence risks in second and third deliveries were high. A history of OASIS had little or no impact on the rates of subsequent deliveries. Women with previous OASIS were delivered more frequently by planned caesarean delivery.

Introduction

Obstetric anal sphincter injuries (OASIS) occur in 1–10% of vaginal deliveries,1–3 and can result in complications such as perineal pain, dyspareunia, as well as urinary and fecal incontinence.4–6 Cross-sectional studies have identified strong risk factors such as primiparity, high birthweight and instrumental delivery,3,7 whereas longitudinal studies have reported that OASIS tends to recur in subsequent births,8–11 but not consistently.12,13

An assessment of the reproductive history in women who have sustained OASIS, focusing on recurrence as well as the likelihood of having a subsequent pregnancy, would be of particular value in counselling women who have had OASIS.

Hospital-based studies on the recurrence of OASIS may be affected by selection bias and small sample size. Additionally, the choice of reference group may have caused some of the inconsistency in reported relative risks of recurrence.12,13 In a big population-based study, Spydslaug et al.9 reported a 4.3-fold increased recurrence risk of OASIS in the second delivery. They also reported an increasing absolute risk of the recurrence of OASIS according to the birthweight of the offspring (23.3% recurrence rate for a birthweight >5000 g). However, little is known about other possible risk factors for recurrence, such as instrumental delivery, interdelivery interval and maternal age. Furthermore, paternal influence, represented in the fetus by characteristics such as birthweight,14 has to our knowledge never been analysed. Additionally, little is known about the recurrence of OASIS beyond the second birth.

One may expect that a severe delivery complication such as OASIS would deter a woman from having a subsequent pregnancy. Previous studies have focused on the quality of life after OASIS,5,15,16 but subsequent delivery rate and the mode of subsequent deliveries has not been properly addressed.11

The aim of the present study was to assess the recurrence risk of OASIS in second and third deliveries, and to study the effect of instrumental delivery, interdelivery interval, maternal age and size of maternity unit on the recurrence of OASIS. We also wanted to estimate the proportion of OASIS cases attributable to a history of OASIS, and to assess the paternal contribution to the recurrence of OASIS. Finally, we wanted to assess the likelihood of having a further delivery and mode of delivery after OASIS.

Methods

In this registry-based cohort study, we used data from the Medical Birth Registry of Norway, which, based on compulsory notification of all live births and stillbirths in the country after 16 weeks of gestation, comprises records of more than 2 000 000 births from 1967 to 2004. A notification form including data on maternal health before and during pregnancy, interventions and complications during delivery and health of the newborn is completed by the midwives and attending physicians. The notification form remained almost unchanged until 1999, when a revised version was introduced.17

All births of a mother were linked by the national identification number, providing sibship files with the mother as the unit of analysis. The analysis was based on mothers with singleton, vertex-presenting infants, weighing 500 g or more, who had their first delivery after 1967: 828 864 mothers in total. In order to compare subsequent rates of OASIS after vaginal births with and without OASIS, women with caesarean in previous deliveries were excluded. Only current vaginal deliveries were followed with respect to the recurrence of OASIS. When subsequent delivery rates from first to second and second to third births were calculated, mothers with caesarean deliveries in previous births (first and first or second, respectively) were excluded, because caesarean delivery may influence further delivery rates.18,19

The main outcome, OASIS, was classified according to the international classification of diseases and included third-degree injury (ICD-10: O70.2), involving sphincter muscle, and fourth-degree injury (ICD-10: O70.3), involving sphincter muscle and rectal mucosa. From 1967 to 1998, OASIS was reported to the Medical Birth Registry as plain text, whereas from 1999 onwards it was reported by checking a box in the form. The registration of OASIS in the Medical Birth Registry of Norway has been validated with a satisfactory result.20,21

The recurrence rate of OASIS was estimated in the mother’s second and third delivery. The odds ratio (OR) of recurrence was defined as the odds of OASIS among women having already had OASIS relative to the odds of OASIS in those without previous OASIS. Adjusted ORs were obtained from logistic regression with adjustment for year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004), instrumental delivery (yes or no), maternal age (<20, 20–29, 30–34, 35–39 and 40 years or older), birthweight (<3000, 3000–3499, 3500–3999, 4000–4499, 4500–4999 and 5000 g or more) and size of maternity unit (<50, 50–499, 500–999, 1000–1999, 2000–2999 and 3000 deliveries per year or more) in the subsequent delivery. The associations of recurrence of OASIS in the subsequent delivery with maternal age, instrumental delivery, birthweight, size of maternity unit and interdelivery interval (<5, 5–9 and 10 years or more) were assessed by logistic regression, restricting these analyses to women with a history of OASIS. To assess whether the effects of birthweight on OASIS in instrumental and non-instrumental delivery were significantly different, an interaction term between birthweight and instrumental delivery was added to the regression model.

In order to increase sample size in analyses of paternal contribution to the recurrence of OASIS, 48 392 pairs of first to second, second to third, third to fourth and fourth to fifth singleton, vertex-presenting vaginal deliveries, with birthweights of 500 g or more, with the same father and different mothers were identified. Among these pairs of births, 18 579 (from 11 372 fathers) and 29 813 (from 17 986 fathers) took place in the same and different maternity units, respectively. To avoid underestimated standard errors caused by the nested structure of the data (one or more pairs of births in the same father), we used multilevel logistic regression analysis.22

To estimate the proportions of all cases of OASIS in the second and third delivery attributable to a history of OASIS, population-attributable risk percentages were estimated as 100 × (incidence in the population–incidence in the non-exposed group)/incidence in the population, on the assumption of a causal relationship between an initial and a subsequent OASIS. Exposed third deliveries were those with either OASIS in the first or second delivery.

The subsequent delivery rate was defined as the percentage of all women who had a delivery (second or third) subsequent to the first or second delivery. The proportion of planned caesareans among all subsequent deliveries was calculated, irrespective of how the delivery was performed (planned or emergency caesarean or vaginal delivery). The classification of caesarean deliveries into emergency and planned caesareans in the Medical Birth Registry was introduced in 1988. Consequently, analyses of planned caesarean deliveries were restricted to the period 1988–2004. In the calculation of the subsequent total delivery rate after the first and second delivery, each woman was observed until the end of the observation period (31 December 2004).

Adjustments were made in a Cox proportional hazards regression of time from OASIS to a subsequent delivery for possible confounding factors in the previous delivery: infant death within 1 year (yes or no); year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004); instrumental delivery (yes or no); maternal age (<20, 20–29, 30–34, 35–39 and 40 years or older); maternal marital status (married, cohabiting, unmarried or single, other, unknown); and maternal level of education (<8, 8–10, 11–12, 13–17, 18 or more years, unknown). Because the Medical Birth Registry covers all births in Norway, lost to follow-up were women who emigrated or died. By logistic regression adjusted for year of delivery we found no significant differences in emigration (0.3–1.6%), nor in maternal death (0.5–1.7%), between groups of women with OASIS or not in first and second deliveries. Data on women who did not have a subsequent delivery were treated as censored observations, with censored time equal to the last date of registration (31 December 2004), the date of emigration or maternal death.

The statistical analyses were carried out in spss (SPSS Inc., Chicago, IL, USA) and mlwin (Centre for Multilevel Modelling, University of Bristol, UK). The regional committee for medical research ethics approved the study protocol (REK Vest no. 247.09).

Results

Figure 1 shows the study population according to the mode of delivery and occurrence of OASIS in first, second and third deliveries. OASIS occurred in 2.8, 1.1 and 0.7% of first, second and third vaginal deliveries, respectively. The data from all 828 864 women, without exclusions, is included in Figure 1.

Figure 1.

 Study population according to mode of delivery and history of obstetric anal sphincter injuries (OASIS) in the first, second and third delivery.

Recurrence of OASIS in second and third deliveries

The occurrence of OASIS in second deliveries subsequent to deliveries with OASIS was 5.6% (750/13 305) and without OASIS was 0.8% (4546/545 469) [OR 4.2; 95% CI 3.9–4.5; relative to women without a history of OASIS, adjusted for year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004), birthweight (<3000, 3000–3499, 3500–3999, 4000–4499, 4500–4499 and 5000 g or more), instrumental delivery (yes or no) and size of maternity unit (<50, 50–499, 500–999, 1000–1999, 2000–2999 and 3000 deliveries per year or more) in the second delivery]. Additionally, forceps deliveries, birthweights >3500 g and maternity units with over 3000 deliveries per year were associated with the recurrence of OASIS in the second delivery (Table 1). Vacuum deliveries only marginally increased the risk of recurrence (OR 1.5; 95% CI 1.0–2.3; Table 1). Maternal age of <40 years and interdelivery interval were not associated with an excess recurrence risk of OASIS (Table 1). Instrumental delivery did not increase birthweight-specific recurrence risks (Table 2). An interaction term between birthweight and instrumental delivery added to the model was not significant (= 0.6).

Table 1.   Risk factors for recurrence of obstetric anal sphincter injuries (OASIS) in the second vaginal delivery in 13 305 women with OASIS in the first delivery, Norway, 1967–2004
CharacteristicsTotal number of deliveriesNumbers of OASIS N (%)Adjusted OR (95% CI)
  1. Adjusted for year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004), birthweight (<3000, 3000–3499, 3500–3999, 4000–4499, 4500–4499 and 5000 g or more), instrumental delivery (yes or no) and size of maternity unit (<50, 50–499, 500–999, 1000–1999, 2000–2999 and 3000 deliveries per year or more) in the second delivery.

  2. *Adjusted for year of delivery and maternal age in the second delivery.

Birthweight(g)
<300084118 (2.1)0.6 (0.4–1.1)
3000–34993167102 (3.2)Reference
3500–39995247252 (4.8)1.5 (1.2–1.9)
4000–44993117249 (8.0)2.5 (1.9–3.1)
4500–4999812104 (12.8)4.2 (3.1–5.6)
5000 or more12125 (20.7)7.1 (4.3–11.6)
Instrumental delivery
Non instrumental12 921702 (5.4)Reference
Vacuum26830 (11.2)1.5 (1.0–2.3)
Forceps11618 (15.5)3.2 (1.9–5.4)
Maternal age(years)
<20651 (1.5)0.5 (0.1–3.6)
20–297698383 (5.0)Reference
30–344387288 (6.6)1.0 (0.9–1.2)
35–39103966 (6.4)1.0 (0.7–1.3)
40 or older11612 (10.3)1.8 (1.0–3.4)
Interdelivery interval(years)*
<511 259623 (5.5)Reference
5–9.91818116 (6.4)1.1 (0.9–1.4)
10 or more22811 (4.8)0.7 (0.4–1.4)
Maternity unit(deliveries/year)
<50460 
50–499117246 (3.9)0.8 (0.6–1.2)
500–999137641 (3.0)0.6 (0.4–0.9)
1000–19993220161 (5.0)Reference
2000–29991959113 (5.8)1.2 (0.9–1.5)
3000 or more5102357 (7.0)1.4 (1.2–1.8)
Outside maternity unit43032 (7.4)1.1 (0.7–1.6)
Table 2.   Recurrence of obstetric anal sphincter injuries (OASIS) in the second vaginal delivery by birthweight and instrumental delivery, Norway, 1967–2004
Instrumental delivery in the second deliveryBirthweight (g) in second deliveryTotal numbers of second deliveriesOASIS in second delivery no. (%)Adjusted OR (95% CI)
  1. Adjusted for year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004), maternal age (<20, 20–29, 30–34, 35–39 and 40 years or older) and size of maternity unit (<50, 50–499, 500–999, 1000–1999, 2000–2999 and 3000 deliveries per year or more) in the second delivery.

No instrumental delivery<300081417 (2.1)0.7 (0.4–1.1)
3000–3499310896 (3.1)Reference
3500–39995129240 (4.7)1.5 (1.2–1.9)
4000–44992991230 (7.7)2.5 (2.0–3.2)
4500–499976196 (12.6)4.4 (3.3–6.0)
5000 or more10722 (20.6)7.9 (4.7–13.3)
Instrumental delivery Forceps/vacuum<3000271 (3.7)1.3 (0.2–9.8)
3000–3499596 (10.2)3.3 (1.4–8.1)
3500–399911812 (10.2)3.3 (1.7–6.2)
4000–449912619 (15.1)5.1 (3.0–8.7)
4500–4999518 (15.7)5.2 (2.3–11.5)
5000 or more143 (21.4)8.0 (2.1–30.0)

A history of OASIS in the first or second delivery increased the occurrence in the third delivery (Table 3). The ORs relative to women without OASIS in the first and second delivery were highest in women with no OASIS in the first delivery but with OASIS in the second delivery, and women with OASIS in both first and second deliveries (adjusted ORs 9.3 and 10.6, respectively; 95% CIs 7.3–11.8 and 6.2–18.1, respectively).

Table 3.   Recurrence of obstetric anal sphincter injuries (OASIS) in third vaginal deliveries, Norway, 1967–2004
OASIS in first deliveryOASIS in second deliveryTotal number of third deliveriesOASIS in third deliveries no. (%)Adjusted OR (95% CI)
  1. Adjusted for year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004), maternal age (<20, 20–29, 30–34, 35–39 and 40 years or older), instrumental delivery (yes or no), birthweight (<3000, 3000–3499, 3500–3999, 4000–4499 and 5000 g or more) and size of maternity unit (<50, 50–499, 500–999, 1000–1999, 2000–2999 and 3000 deliveries per year or more) in the third delivery.

No OASISNo OASIS207 2991106 (0.5)Reference
No OASISOASIS117982 (7.0)9.3 (7.3–11.8)
OASISNo OASIS3986124 (3.1)4.0 (3.3–4.9)
OASISOASIS16916 (9.5)10.6 (6.2–18.1)

The population-attributable risk percentage of OASIS in second and third deliveries as a result of previous OASIS was 10 and 15%, respectively.

No time trends of ORs of recurrence in second or third deliveries were found, as assessed by adding an interaction term between year of the current delivery and a history of OASIS to the model or stratifying for year of birth (not presented).

A man who fathered a birth resulting in OASIS was more likely to father a subsequent birth resulting in OASIS in another woman who gave birth in the same maternity unit (adjusted OR 2.1, relative to men with no history of OASIS, for year of delivery, maternal age and maternal birth order in the current delivery; 95% CI 1.2–3.7; 5.6%, compared with 2.3%). Adjusting for birthweight had a negligible effect. However, if the deliveries took place in different maternity units, the recurrence risk was not significantly increased (OR 1.3; 95% CI 0.8–2.1; 4.4% compared with 2.9%).

Subsequent delivery after OASIS

After OASIS in the first delivery, 66.7% of women had a second delivery compared with 76.9% of women with no OASIS in the first delivery (Table 4). However, adjusted hazard ratios revealed no significant differences. Consistently, after stratifying analyses by year of first delivery, no significant differences were observed in second delivery rates between women with and without OASIS in the first delivery (not presented). Cumulative proportions by time from deliveries with and without OASIS were almost the same (not presented). Women with OASIS in the first delivery more frequently had a planned caesarean in the second delivery than women without OASIS in the first delivery (6.0 and 1.5%, respectively; Table 4). The difference persisted after adjustment for possible confounding factors (OR 3.0; 95% CI 2.8–3.3).

Table 4.   Subsequent pregnancy after first vaginal delivery with or without obstetric anal sphincter injuries (OASIS), total and planned caesarean in second delivery, Norway, 1967–2004
First deliverySecond delivery
 Total number of women with first vaginal deliveries no. (%)Total number of second deliveries no. (%) Adjusted hazard ratio (95% CI)Numbers of planned caesarean in second delivery no. (%)* Adjusted OR (95% CI)
  1. Women delivered by caesarean in the first delivery were excluded.

  2. Hazard ratios and odds ratios are adjusted for infant death within 1 year (yes or no), year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998, and 1999–2004), instrumental delivery (yes or no), maternal age (<20, 20–29, 30–34, 35–39 and 40 years or older), maternal marital status (married, cohabiting, unmarried or single, other, unknown) and maternal level of education (<8, 8–10, 11–12, 13–17, 18 or more years, unknown) in the first delivery.

  3. *Numbers of second deliveries after 1988 without and with OASIS (denominators) were 271 512 and 11 065, respectively.

No OASIS734 245 (100)564 826 (76.9)4050 (1.5)
ReferenceReference
OASIS21 676 (100)14 461 (66.7)658 (6.0)
1.02 (1.00–1.04)3.0 (2.8–3.3)

Women with OASIS in the first or second delivery had a lower subsequent delivery rate (from the second to the third delivery) than women without a history of OASIS (Table 5). However, adjusted hazard ratios revealed small or insignificant differences. Consistently, after stratifying analyses by year of first delivery, no significant differences were observed in third delivery rates between women with and without OASIS in first or second delivery (not presented). Women with a history of OASIS in the first or second delivery were delivered significantly more frequently by planned caesarean in the third delivery (Table 5). Cumulative proportions by time from deliveries with and without OASIS were almost the same (not presented). In women with OASIS in both first and second deliveries, the rate of planned caesarean delivery was 13 times higher than in women with no previous OASIS (adjusted OR 13.4; 95% CI 9.1–19.7) (Table 5).

Table 5.   Subsequent pregnancy after two vaginal deliveries with or without obstetric anal sphincter injuries (OASIS), total and planned caesarean third deliveries, Norway, 1967–2004
First deliverySecond delivery Third delivery
  Total number of women with two vaginal deliveries N (%)Total numbers of deliveries N (%) Adjusted hazard ratio (95% CI)Numbers of planned caesarean N (%)* Adjusted OR (95% CI)
  1. Women delivered by caesarean in the first and second delivery were excluded.

  2. Hazard ratios and odds ratios are adjusted for infant death within 1 year (yes or no), year of delivery (1967–1974, 1975–1982, 1983–1990, 1991–1998 and 1999–2004), instrumental delivery (yes or no), maternal age (<20, 20–29, 30–34, 35–39 and 40 years or older), maternal marital status (married, cohabiting, unmarried or single, other, unknown) and maternal level of education (<8, 8–10, 11–12, 13–17, 18 or more years, unknown) in the second delivery.

  3. *Numbers of third deliveries after 1988; denominators in the same order according to OASIS in the first two deliveries were 120 907, 1099, 3348 and 201, respectively.

No OASISNo OASIS540 923 (100)215 823 (39.9)1552 (1.3)
ReferenceReference
No OASISOASIS4546 (100)1331 (29.3)92 (8.4)
0.90 (0.85–0.95)6.2 (4.9–7.7)
OASISNo OASIS12 555 (100)4181 (33.3)74 (2.2)
1.01 (0.98–1.04)1.7 (1.3–2.1)
OASISOASIS750 (100)213 (28.4)34 (16.9)
1.08 (0.94–1.23)13.4 (9.1–19.7)

Discussion

Women with a history of OASIS in the first and the two-first deliveries had four- and ten-fold increased risks of OASIS in the subsequent delivery, respectively. The recurrence of OASIS was strongly associated with forceps delivery and birthweights of 3500 g or more in the second delivery. However, instrumental delivery did not further increase the excess recurrence risk observed in heavy newborns. The recurrence risk of OASIS increased with size of maternity unit, but was not influenced by maternal age under 40 years or by the interdelivery interval. A man who fathered a birth that resulted in OASIS was more likely to father a subsequent birth that resulted in OASIS in another woman. However, if the deliveries took place in different maternity units, the recurrence risk was not significantly increased. The subsequent delivery rate was not different in women with and without previous OASIS, whereas women with previous OASIS were more often scheduled for caesarean delivery.

Strengths of our study include the population-based design and the prospective collection of data, reducing the risk of selection and recall bias. The national identification number allowed the linkage of births to both parents. The long follow-up period allowed us to study recurrence risk and subsequent delivery rates beyond the second delivery. Our nationwide registry allowed for the collection of data in subsequent deliveries, regardless of any change in maternity unit. The OASIS variable has a high validity.20,21 However, increased awareness in the current pregnancy may have caused a higher sensitivity or case ascertainment in a delivery after previous OASIS. Furthermore, data on several possible confounding factors were available.

In this study, women who had OASIS in the first delivery had a roughly four-fold excess risk of OASIS in the second delivery. Although the risk of the occurrence of OASIS tends to decrease with vaginal birth order,3 we generally noted an increased recurrence risk in birth order 3. Women who had OASIS in the second delivery, but not in the first, were nine times more likely to have OASIS in the third delivery, and also had high absolute risks (7%). The risk in women who had OASIS in the two-first deliveries was particularly high (absolute risk 9.5%, adjusted OR 10.6). Also, women who had OASIS in the first, but not the second, delivery were at increased risk in the third delivery, although with a modest absolute risk (3.1%). This could not result from the exclusion of women with caesarean delivery in the second delivery, because in a supplementary analysis, inclusion of caesarean delivery had little effect on ORs. The high absolute risk in the third delivery in women with OASIS in the second delivery was not expected, and justifies attention to the recurrence of OASIS beyond the second delivery. It cannot be ruled out that women with a history of OASIS and subsequent caesarean delivery more often had fourth-degree OASIS and therefore higher recurrence risks if they were delivered vaginally.

To our knowledge, nine studies have previously reported on the recurrence risk of OASIS,8–13,23–25 with conflicting results. Seven of these studies have reported increased recurrence risk of OASIS in a second delivery,8–11,23–25 and another two studies found no increased risk of OASIS in women with prior OASIS.12,13 The last two studies included women with OASIS in birth order 1 in the reference group, and thus probably underestimated the relative risks arising from the higher risk of OASIS in the first deliveries, concluding that prior OASIS does not increase recurrence, and that the increased recurrence found in previous studies could be caused by bias.

Although most women sustaining OASIS in second and third deliveries did not have previous OASIS, as many as 10% of all cases of OASIS in the second delivery, and 15% of cases of OASIS in the third delivery, were attributable to a history of OASIS.

Consistent with another study based on the Medical Birth Registry of Norway,9 birthweights of 3500 g or more were strongly associated with a recurrence of OASIS. This indicates that for women with a history of OASIS and high estimated fetal weight, caesarean delivery must be considered.9

As in previous studies,12,13 in the present study instrumental deliveries, and particularly forceps deliveries, were strongly associated with a recurrence risk of OASIS. Vacuum deliveries only marginally increased the recurrence risk of OASIS. Therefore, vacuum delivery is probably a better choice than forceps when women with prior OASIS are delivered instrumentally, unless the clinician is skilled in forceps delivery.

Although instrumental delivery was strongly associated with the recurrence of OASIS, it did not further increase the excess recurrence risk in heavy newborns. This may be useful information for the clinical decision of whether an instrumental delivery of a large infant should be performed in a woman with a history of OASIS.

We have previously reported that maternal age is associated with the occurrence of OASIS.3 However, our results indicate no association of either maternal age under 40 years or interdelivery interval with the recurrence of OASIS. Our results provide reassurance that recurrence risk in older women is not substantially different from that in younger women, and that the time to the next pregnancy does not seem to influence recurrence. The recurrence risk of OASIS was higher in the maternity units with more than 3000 deliveries per year. After adjusting for instrumental delivery, which is more common in referral hospitals, the higher risk persisted. However, it cannot be ruled out that the excess risk was the result of better registration, diagnostic skills or referral of complicated pregnancies to larger maternity units.

Most risk factors for OASIS relate to the mother, and little is known about a potential paternal influence on OASIS. A change of female partner after a birth with OASIS should remove the previous mother’s genetic contribution to the recurrence risk. The excess paternal recurrence rate was not present if both deliveries took place in different maternity units, which contradicts a biological paternal effect. However, a man who fathered a birth with OASIS was more likely to father a subsequent birth with OASIS in another woman if the delivery took place in the same delivery unit. Only including births that took place in the same delivery unit would hold effects such as practices of perineal protection more constant. This potential genetic paternal effect warrants further study.

Some studies on the quality of life after OASIS have reported conflicting results,5,15,16 but subsequent fertility and mode of subsequent deliveries have not been emphasised. Consistent with the present study, one study has reported a reduced unadjusted likelihood of having a further delivery after OASIS.11 After stratification by year of first delivery in the present study, no differences from expected values were generally observed in subsequent delivery rates (data not shown). A recent small case–control study showed that women who had OASIS wished to postpone or avoid a further delivery, but consistent with the present study a history of OASIS did not influence further delivery rates.26 However, the excess rate of planned caesarean delivery in the present study, which has been reported to be associated with increased risk of severe maternal and neonatal morbidity and mortality,27 persisted.

Conclusion

The absolute and relative recurrence rates after OASIS in the first and second deliveries were high. Therefore, emphasis should be placed on counselling women after an initial OASIS, and attention should be paid to prevent OASIS in the first delivery. A history of OASIS had little or no impact on the subsequent delivery rate. However, women with previous OASIS more frequently had planned caesarean delivery. The potential risk factors related to the father should be further studied.

Disclosure of interests

All authors declare that they have no relevant interests to declare.

Contribution to authorship

EB contributed by writing the article, performing the statistical analyses, in the conception and design of the study, and in the interpretation of data. LMI contributed by supervising, drafting the article and revising it for important intellectual content. PEB contributed by supervising and drafting the article, and by revising it for important intellectual content. The main supervisor, SR contributed by revising the article, by the conception and design of the study, by the interpretation of data and by supervising the statistical analyses. All authors approved the final version of the article.

Details of ethics approval

The regional committee for medical research ethics approved the study protocol (REK Vest no. 247.09).

Funding

The study was funded by the Norwegian Foundation for Health and Rehabilitation and the Norwegian Women’s Public Health association.

Acknowledgements

The authors want to thank the Norwegian Foundation for Health and Rehabilitation and the Norwegian Women’s Public Health association for funding the study.

Ancillary