Our results are consistent with other studies that show that pregnancies in women with spinal cord injury are high risk but pregnancy outcomes are generally good . The obstetric management of women with SCI in our unit are summarised elsewhere and will not be repeated here except where relevant to the findings of the study .
We found a high rate of antenatal complications, particularly worsening of spasms (38%) which did not require any further treatment and urinary tract infections (24%). This is consistent with other studies suggesting a higher incidence of UTIs when compared with the general population  and supports the practice of frequent monitoring with consideration of antibiotic prophylaxis . Fetal malpresentation at term, including breech and transverse lie, was more common with 13.6% incidence, approximately five times that of the general obstetric population at 3% , with a higher incidence of 21% in women with lesions above T6 affecting the abdominal myotomes. A possible mechanism for this increased rate of malpresentation could be the increased laxity of abdominal wall muscles in women with higher spinal cord injury. The primary indication for Caesarean section was malpresentation in five cases, with malpresentation co-existent in the other three cases where elective Caesarean was performed for alternative indications. In this study, we also report successful external cephalic version in one patient with a T11 injury with no adverse effects.
Women with SCI do not appear to be at higher risk of developing hypertensive disorders of pregnancy, with an incidence of 4.5% comparable to the national incidence of between 5 and 7% . The incidence of venous thromboembolism (VTE) was higher at 4.5% than the background incidence in the general obstetric population of 0.2% however this result should be interpreted with caution as the three women in our study with proven VTE were known to have other pre-disposing risk factors such as SCI during pregnancy and high BMI of more than 50. Our practice is to perform standard VTE risk assessment in line with national guidance . Pressure ulcers were less frequently encountered than expected as an antenatal complication occurring in 3% of pregnancies. Reported rates of pressure ulcers in the general spinal cord injury patient population are as high as 25 to 66%  and our findings may reflect pregnancy as a marker of good function and health.
Our study shows a prevalence of preterm birth at 18% which is higher than the estimated UK national rate of 5–9%  but lower than the reported rates in the literature regarding SCI patients of up to 27% . Urinary tract infections are associated with an increased risk of preterm birth and in our study population, 33% of the women who had a preterm birth having also had a confirmed urinary tract infection during pregnancy. Women with lesions at and above T10 are at risk of not perceiving contractions and our data suggests that higher level spinal injuries are associated with a higher incidence of preterm labour of 25% in women with lesions above T10 and 4.5% in women with a lesion below T10. Reassuringly the only extremely preterm delivery at 30 weeks was iatrogenic for maternal urosepsis and the remainder occurred between 35 and 37 weeks of gestation.
Neonatal outcomes were good with a very low incidence of neonatal admission to NICU. The prevalence of small for gestational age babies below the 10th percentile by customised growth chart in our population was 10% which is comparable to the national incidence of 8% . This is reassuring but the study is insufficiently large to detect any significant difference or to make any recommendations regarding monitoring of fetal growth and we would therefore recommend an individualised approach.
Our rate of Caesarean section at 23% is lower than reported in other studies of women with SCI, where rates of up to 69% are described . This compares favourably with the Caesarean section rate in the general UK obstetric population which is reported as 27.1% in 2015–2016 . The rate of Caesarean section in the general population has increased over the 25 year period covered by our study and this was mirrored in women with SCI but the relatively small numbers of patients delivering annually in our study cohort mean that further analysis of this trend is not possible. One qualitative study from Switzerland described some women with SCI reporting a lack of choice with regard to mode of delivery due to their medical providers’ lack of familiarity with their condition . The rate of intrapartum autonomic dysreflexia was 6%. Of these four women, all had inadequate analgesia at the time of dysreflexia. Three had vaginal deliveries after successful regional anaethesia and one required an emergency Caesarean under spinal anaesthesia in order to achieve resolution of dysreflexia. We attribute this low rate of intrapartum autonomic dysreflexia to our protocol of early use of epidural anaesthesia in labour for women at risk of this complication.
All of the women over the 25-year period of our study were under the care of a designated consultant obstetrician, with an interest in spinal cord injury working at the same site as the National Spinal Injuries Centre, and specialised consultant, nursing, midwifery and anaesthetic staff familiar and confident in spinal cord injury management. Women received the majority of their antenatal care as outpatients in their local hospitals with late pregnancy and intrapartum management as inpatients in our hospital. We suggest that the relatively high numbers of SCI pregnancies managed in this unit may have contributed to this lower rate of Caesarean through experienced consistent antenatal and intrapartum management. The majority of women in this study successfully and safely achieved vaginal delivery regardless of the level of spinal cord injury and our data is reassuring in supporting patient choice with regard to mode of delivery.
This study reports pregnancy outcomes in a small population of high risk women with SCI and therefore it is difficult to generalise practice and derive meaningful comparisons between groups of women in this cohort however, this is the largest recent case series study since 1972 and may help counsel women regarding likely outcomes. Possible ascertainment bias is also a potential limitation given the retrospective nature of the study, however due to the hospital specific practice of referring all pregnant patients with spinal cord injury to a single designated obstetrician, it is unlikely that relevant cases were not included in the study. The length of the study over 25 years is a strength but may limit the generalizability of the outcomes, since some aspects of obstetric care, such as the rate of Caesarean section, have increased over time.