Maternal origin matters: Country of birth as a risk factor for obstetric anal sphincter injuries

Obstetric anal sphincter injuries (OASIS) are severe complications to vaginal births with potential long‐term consequences. Maternal origin has been proposed to affect the overall risk, but the association and underlying explanation are uncertain. The objective was to assess the association between maternal country of birth and OASIS.


| INTRODUC TI ON
Obstetric anal sphincter injuries (OASIS) are severe complications to vaginal deliveries associated with both short-and long-term consequences.5][6] However, the incidence varies greatly both between different populations and over different time periods.
Several previous studies have investigated the relationship between ethnicity, race, country of birth, and OASIS.How ethnicity or race is defined varies between different studies.In our previous study 7 we found an increased risk of OASIS in those born outside of Europe which we wanted to explore further.Asian ethnicity has been associated with a significantly increased risk of OASIS, [8][9][10] and women from South Asia have been identified as being particularly high risk. 113][14] Short stature has been associated with poor perinatal outcomes, and more significantly so in some ethnic groups, 15 but the role of maternal height in the overall risk of OASIS and how this relates to maternal origin have, to our knowledge, only been addressed in one previous study. 16This relationship needs to be further evaluated to understand the underlying mechanisms.
Women migrating from low-income countries are at higher risk of suboptimal care and poor maternal outcomes. 17Women from Sub-Saharan Africa living in Nordic countries have been identified as an obstetric high-risk group 18,19 and a higher risk of OASIS has also been observed. 20Female genital circumcision (FGC) has also been proposed to modulate the risk of OASIS.[23] It is of great importance to investigate and understand what role maternal origin plays in obstetric outcomes, considering worldwide increasing migration.
Our objective in this study was to explore how maternal country of birth is associated with risk of OASIS in the Swedish population, using the nationwide mandatory obstetric register, and also to looki at what role maternal stature plays.Secondarily, we wanted to investigate the association between FGC and OASIS.

| MATERIAL S AND ME THODS
In this population-based study, data was extracted from the Swedish Medical Birth Registry, including all term (≥37+0 weeks) singleton vaginal deliveries between 2005 and 2016.The information in the registry is collected prospectively from standardized antenatal, obstetric, and neonatal records at all antenatal clinics and hospitals in Sweden.The Medical Birth Registry is validated and contains demographic data, reproductive history, maternal diseases, and pregnancy and neonatal outcomes of more than 98% of all births in Sweden. 24Cases of OASIS were identified through checkboxes indicating injury to the sphincter or rectum in the standardized obstetric records, or ICD-codes O702 and O703 reported to the Medical Birth Registry or the National Patient Registry within 6 weeks of delivery.OASIS is diagnosed as thirdor fourth-degree perineal lacerations according to classification by Sultan et al. and diagnosis is done by either midwives or doctors at the time of suturing. 25formation regarding maternal educational level and country of birth was obtained through linkage with Statistics Sweden.No direct patient involvement was needed and therefore no informed consent required.
Maternal country of birth was divided into categories according to groupings used by the National Board of Health and Welfare.A list of all countries in each category is provided in Table S1.

| Statistics
Chi-squared analyses were performed to assess the differences regarding maternal characteristics by maternal country of birth.We applied modified Poisson regression for multivariate analyses to examine the association between maternal country of birth and OASIS.
Results are presented as risk ratios (RRs) and adjusted RRs (aRRs) with 95% confidence intervals (CIs).A P-value less than 0.05 was considered statistically significant.
In the adjusted models, we first added maternal characteristics (age, level of education, smoking, year of delivery, body mass index [calculated as weight in kilograms divided by the square of height in meters], and parity) as covariates.To specifically explore the role maternal height plays in risk of OASIS, this was added to the previous covariates in our second analysis.Thirdly, we added risk factors associated with delivery-episiotomy and instrumental delivery.
Lastly, we also adjusted for birth weight which was added to the previously mentioned variables.
Missing data for the variables maternal educational level, body mass index, height, age, and smoking were replaced by the overall means.Missing data in all variables was less than 10%.
A stratified analysis comparing primiparous and multiparous women was performed.For multiparous women, adjustments were done for previous cesarean section and for parity >2.Chi-squared tests were performed to test for homogeneity of the aRRs across investigated strata, in which the log RRs were weighted according to precision (1/variance of the log aRR).
Sensitivity analyses were performed to explore the association between female genital FGC and OASIS.Among women from Sub-Saharan Africa, those who had undergone FGC were compared with those who had not.Furthermore, the effect of episiotomy among women who had undergone FGC was investigated.

| Ethics statement
The study was approved by the Regional Medical Ethics Committee, Lund, Sweden (2018/538).

| RE SULTS
Between 2005 and 2016, 988 804 singleton vaginal deliveries were included.Table 1 shows descriptive maternal characteristics by maternal country of origin.The differences were all statistically significant, with P values <0.001.Women from Sub-Saharan Africa had the highest proportion of multipara (49%), the highest proportion of obesity (23%), and a low level of maternal education.They also had the highest rates of episiotomy (11.4%) and previous cesarean section (6.5%).Women from East/Southeast Asia had the highest proportion of underweight women (9%) and were significantly shorter than Swedish-born women.The highest rate of instrumental deliveries was found among women from East/Southeast Asia.
The overall rate of OASIS was 3.5% and remained constant over the study period.Table 2 shows the risk of OASIS by parity and maternal country of birth.The highest risk of OASIS both among primiparas and multiparas was found in women from East/Southeast Asia (10.6% in primiparous and 3.9% in multiparous women).Primiparous women from Sub-Saharan Africa had a higher risk of OASIS as compared with Swedish-born women (9.1% vs. 5.8%), whereas women from South/Central America had a lower risk (4.7%).
Table 3 shows the crude and adjusted RRs for OASIS by maternal country of birth.Women from the other Nordic countries had a slightly higher risk of OASIS than Swedish-born women.Among women from Sub-Saharan Africa, the risk of OASIS was significantly increased after adjusting for maternal characteristics, height, and delivery mode.The risk was even further increased (aRR 1.87) when adjusting for birth weight.Women from the Middle East had a slightly increased risk of OASIS (aRR 1.09), but only after adjustment through all four steps.Women from East/Southeast Asia had a significantly increased risk of OASIS, which was less pronounced, but still profoundly elevated, after adjusting for maternal height and delivery mode.When adjusting for birth weight, the risk increased slightly again.Among women from South/Central America, the risk of OASIS was significantly decreased and the calculations only marginally changed through the adjustments.
Table 4 shows a stratified analysis of primiparous and multiparous women.For women from East/Southeast Asia, the increased risk of OASIS was significantly more pronounced among multiparous women (aRR 2.41) than among primiparous women (aRR 1.67) (P for homogeneity over parity strata <0.001).For women from Europe/ USA/Canada/Australia/New Zealand, the RR of OASIS was reduced for primiparous women (aRR 0.90) and increased for multiparous women (aRR 1.20) as compared with their Swedish counterparts (P for homogeneity over parity strata <0.001).Among women from South/Central America, the risk reduction was most pronounced among multiparous women (aRR 0.46).For the remaining three investigated groups, the RRs for OASIS, were of approximately the same magnitude among primiparous and multiparous women.
We then performed sensitivity analyses to investigate what effect FGC has on the risk of OASIS.Among women from Sub-Saharan Africa, 4.4% were reported to have undergone through FGC (n = 927).In this migration group, FGC substantially increased the risk of OASIS (aRR 3.05, 95% CI 2.60-3.58).We then investigated if episiotomy could have had a protective effect on women who had undergone FGC.Among all women with reported FGC (n = 1178), the aRR for episiotomy was 0.73 (95% CI 0.52-1.03).For women from Sub-Saharan Africa without FGC, the aRR for OASIS after episiotomy was 0.95 (95% CI 0.92-0.98).There was no significant statistical difference in comparison between the two point estimates of the protective effect of episiotomy between the women with or without FGC (P = 0.14).

| DISCUSS ION
In women undergoing term, singleton deliveries between 2005 and 2016, we found women born in East/Southeast Asia and Sub-Saharan Africa to be of increased risk of OASIS as compared with women born in Sweden.For women from South/Central America the risk was instead decreased.Among women from Sub-Saharan Africa, FGC markedly increased the risk of OASIS.Episiotomy had an overall protective effect against OASIS, but the effect was not statistically more pronounced among the women that had undergone FGC.
The higher risk of OASIS among women from East/Southeast Asia correlates to previous studies in other populations. 8,9atomical differences have been proposed as a potential explanation.Pelvimetry studies have shown Asian women to have a more horizontal pelvis, theoretically changing the direction of expulsive forces and increasing the risk of perineal damage. 26Racial differences have been found in skin resistance, where Asian skin has been found to be more fragile than Caucasian or African American skin. 27terestingly, low rates of OASIS in Asian countries have been reported by the WHO-0.2% in India, 0.4% in China and 0.9% in Thailand. 28Differences in diagnosis and obstetrical practices might explain some of the discrepancies in rates of OASIS among Asian women giving birth in their home countries compared with Western countries, as seen in a prospective study by Bates et al., comparing   births in Sydney and Hong Kong. 291][32] Somali or Sub-Saharan  Abbreviations: BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters); CS, cesarean section; FGC, female genital circumcision; NZ, New Zealand.

TA B L E 1 (Continued)
African women have also been found to be at risk of adverse obstetric outcomes. 30,33The underlying cause is unknown and we speculate that the risk increase could be associated with language and cultural differences, such as strong preference for vaginal birth, 34 and possibly genetic alterations.Another potential explanation could be a difference in obstetrical care provided.
Women from South/Central America appeared to be at decreased risk of OASIS.One putative explanation could be differences in genetic components of skin and collagen and in body stature, with a difference in hip size in relation to height, although there is no certain evidence available.WHO reported low prevalence of OASIS in Latin American countries-from 0.3% in Cuba to 1.2% in Ecuador. 28wever, differences in obstetrical practice and diagnosis make it difficult to compare incidence rates.
Language skill has been proposed as an independent risk factor for OASIS.Non-native speaking women have been found to have a two-fold increased risk 35,36 and the need for an interpreter has been independently associated with risk of OASIS. 8A recent Norwegian study found that foreign-born women with a Norwegian-born partner have a lower risk of OASIS than those with foreign-born partners. 16They also found that long residence in Norway lowered the risk of OASIS compared with newly migrated women. 16This indicates that improved communication and familiarity with the healthcare system improves outcomes.
A previous meta-analysis concluded that there is sufficient evidence to associate FGC with adverse obstetric outcomes. 37It has also been suggested that obstetric risk increases with extent of infibulation (including cesarean section, postpartum hemorrhage and stillbirth). 38In a recent Swedish study, episiotomy was more common among women with previous FGC and the risk of OASIS was almost three-fold higher than those without FGC. 23It has been proposed that infibulation causing scarring, keloid formation, and retention cysts can decrease the elasticity of the skin and therefore lead to more extensive perineal tearing. 39driguez et al. 40 found episiotomy to be protective of OASIS among women with FGC.This could not be statistically confirmed in our results, although there appeared to be a protective trend.We speculate that the awareness among health professionals of OASIS in women with FGC underlies the increased episiotomy rate.This finding emphasizes the importance of improving obstetric management in high-risk groups, where liberal use of episiotomy could be questioned.
Maternal height has to our knowledge only been used as a covariate in one previous study when looking at risk of OASIS. 16This factor is closely related to maternal origin, and by doing a stepwise analysis we were able to analyze its role separately.When maternal height is added to the adjustments, the risk of OASIS decreases among women from East/Southeast Asia; however, the risk is still significantly increased.Hence the overall risk disparity can only partly be attributed to height differences between ethnic groups, which is an important finding in a clinical setting.
The major strength of this study is the use of a large populationbased registry with validated high-quality data input.This enabled analysis of subgroups with sufficient numbers to yield results of high statistical power.The dataset contained a relatively low percentage of missing values.Limitations include the fact that residual confounding might still be present, and data on length of second stage of labor and use of analgesia were not available.Furthermore, using country of birth as our main outcome measure means there will be some ethnic diversity within the groups, which might lessen the associations.We suspect that there is an under-reporting of FGC considering the incidence of FGC reported by the WHO. 41We also recognize that the diagnosis of OASIS is subjective and relies on the individual examiner.

| CON CLUS ION
Country of birth plays an important role in the risk of OASIS.Women from East/Southeast Asia and Sub-Saharan Africa were at significantly higher risk of OASIS than Swedish-born women.By contrast, women from South/Central America instead had a considerably lower risk.When adjusting for maternal height, the risk associations were only partly altered, indicating other factors underlying the risk differences.FGC was found to be a significant risk factor for OASIS.
Our findings are important in recognizing women at high risk of OASIS in a clinical setting, and future studies should be aimed at understanding the reason for these differences and improving the care for those women at highest risk to ultimately reduce the frequency of OASIS overall.

Crude
Risk of obstetric anal sphincter injuries (OASIS) by parity and maternal country of birth.
TA B L E 2 Risk of obstetric anal sphincter injuries by maternal country of birth, stratified by primiparous and multiparous status.Abbreviations: aRR, adjusted risk ratio; BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters); CI, confidence interval; NZ, New Zealand.a Adjusted for age, BMI, smoking, year of delivery, maternal height, educational level, episiotomy, instrumental delivery, and birth weight.
Abbreviations: aRR, adjusted risk ratio; BMI, body mass index (calculated as weight in kilograms divided by the square of height in meters); CI, confidence interval; NZ, New Zealand.a A = parity, education, age, BMI, smoking, year of delivery.b B = parity, education, age, BMI, Smoking, year of delivery, height.c C = parity, Education, age, BMI, smoking, year of delivery, height, episiotomy, instrumental delivery.d D = parity, education, age, BMI, smoking, year of delivery, height, episiotomy, instrumental delivery, birth weight.TA B L E 4 b Adjusted for above variables, and also parity >2 previous children and previous cesarean section.