Discussion
Our results support the increased risk of OASIS related to well-established risk factors, including primiparity, instrumental delivery and high birth weight. We showed an association with increasing maternal age, obesity (if unadjusted for birth weight), lower maternal height and previous CS. Smoking and low maternal education level were negatively associated with OASIS.
Increasing BMI showed a decreased risk of OASIS in the crude analysis, however only apparent in the adjusted model after adding birth weight to the analysis. Obese women are more likely to give birth to larger babies and it is therefore important to consider whether this apparent protective effect is clinically relevant, or if their overall risk of OASIS is instead increased secondary to risk of foetal macrosomia as seen in the unadjusted results. A decreased risk of OASIS was seen in two previous Swedish studies, with a more profound risk reduction in BMI >35. Constable et al., showed a risk reduction (aOR 0.27) among women with BMI >35 in an Australian setting, and Garetto et al. observed an aOR 0.75 in obese American women. All previously mentioned studies either adjusted for birth weight or excluded cases with babies >4000g. On the contrary, other studies have not been able to statistically detect any protective effect, whereas Landy et al., only found a risk decrease among nulliparous women.
The proposed underlying mechanisms of a protective effect include an increased amount of adipose tissue making the perineal tissue softer and more stretchable, and a larger perineal body increasing the anovaginal distance. These hypotheses are still to be scientifically proven. In contrast, perineal tissue has been suggested to be weaker in obese women; higher BMI has been associated with striae gravidarum which in turn has been linked to a higher risk of perineal trauma. Another consideration is potentially increased missed diagnoses relating to difficulties in examining women with voluminous amount of adipose tissue.
Women with <9 years education had a 15% lower risk of reported OASIS as compared to those with >3 years of higher education after adjusting for maternal characteristics. Similar results are seen in a Finnish study where higher socioeconomic status (SES) was associated with an 18% increased risk of OASIS. A British study showed that women of high SES were more than twice as likely to suffer from OASIS as compared to those of low SES. This difference is somewhat surprising and difficult to explain. It either stands for an unknown confounding factor not accounted for, or potentially by a difference in the provision of health care. The higher risk in women with higher educational level could instead represent a lower risk of missed diagnosis.
We found smoking to be negatively associated with OASIS (cRR 0.49), remaining after adjusting for maternal characteristics, and surprisingly also after adjusting for delivery mode and birth weight (aRR of 0.74). Similar results are seen in a Finnish study where smoking decreased the incidence of OASIS in primiparous women by 28%, and by 20% in a previous Swedish study among primiparous women. The underlying mechanisms of the apparent protective effect are yet unexplained. The results appear paradoxical to other known effects of smoking such as premature skin ageing, reduced blood flow, and an accumulation of body-wide oxidative damage. This causes us to speculate around the true protective effect of smoking, or whether a difference in provision of health care can explain the results.
Maternal age was inversely related to risk of OASIS where women <20 years of age had a 50% lower risk of OASIS compared to women aged 24-29 years. Women >35 years on the other hand, had an approximately 20% risk increase. The highest age group (>40) had a slightly lower risk increase after adjustments than women aged 34-39, but the difference between the risk estimates was far from significant (p=0.20). Gurol-Urgandi et al., showed a similar risk decrease in teenage mothers, but only a slight risk increase (aOR 1.07) in women aged 30-34. A previous Swedish register-based study showed the risk of OASIS to increase almost continuously with age irrespective of parity, with exception for nulliparous women where the risk did not further increase >35 years which was similar to our findings. In this previous study, the aOR was around 2 for nulliparous women >30 years, however the reference group differed from ours (<25 years) which could explain the difference in calculated risks. Similarly, Hornemann et al. found a higher maternal age in women with OASIS compared to women with less severe perineal lacerations. The specific effects of ageing on perineal tissue have been poorly researched and age-specific advantage of protective measures have not been studied. Considering the trends in modern society to start a family at an older age the risk should be acknowledged, and attention should be given to the importance of accurate diagnosis and treatment.
Maternal height was inversely related to risk of OASIS. Women <155 cm had an aRR of 1.34 after adjusting for maternal characteristics. We stipulated adjustments for delivery mode and birth weight would lessen this association, but the correlation instead became more profound and aRR increased to 1.72, suggesting maternal height as a significant independent risk factor. Räisänen et al. found an increase of 1 cm in maternal height to decrease the risk of sphincter injury by 2%. However, there is limited research on the effects of maternal height on obstetric outcomes.
Our results show an increased risk of OASIS among women achieving VBAC compared to primiparous women. D’Souza et al. found a similar risk of OASIS after VBAC compared to that of primiparous women, and a 6.8-fold risk increase in secundiparous women as compared to multiparous women with previous vaginal births only. Räisänen et al. found a 1.4-fold risk increase associated with VBAC and propose an underlying foetopelvic disproportion as a potential explanation, putting women at risk of initial CS and subsequently of OASIS in following vaginal deliveries.
Strengths of this study include using a population-based register providing a large study population available for inclusion. This enabled analysis of multiple risk factors to produce reliable results. The Swedish MBR is of high quality with <2% missing data and is mandatory across the population providing non-selective standardised data. Nevertheless, the register contains some errors and missing values. We also recognise that the classification and diagnosis of OASIS is subjective and depends on the care provider and local routine. Furthermore, our analysis did not include foetal head circumference, length of second stage of labour, use of anaesthetic or episiotomy rates which might impact the overall risk.