Obstetric Considerations in AVM Management

Background

Subarachnoid hemorrhage (SAH) is among the most common nonobstetric causes of maternal mortality during pregnancy and the puerperium (the first 6 weeks after delivery), with around 25% of these hemorrhages being attributed to rupture of intracranial arteriovenous malformations (iAVMs). While the overall incidence of stroke and hemorrhage in pregnancy is rare, rupture of aneurysms and iAVMs can cause devastating neurologic injury to both mother and fetus. The Stroke Council of the American Stroke Association (AHA Scientific Statement) determined that pregnant patients are in a “special consideration” category due to the inconclusive data surrounding AVM hemorrhage and rebleeding risk in pregnancy. Furthermore, pregnant patients are unique in that the safety and well-being of both the pregnant mother and the fetus must be addressed in the medical and surgical management of iAVMs in pregnancy. Within this chapter, we review the available data for potential risks of iAVM hemorrhage and rebleeding in pregnancy, during delivery, and in the postpartum period. Additionally, we review the data regarding the diagnosis and evaluation and selection of operative and nonoperative treatment strategies for the pregnant patient.

Physiologic Changes Associated With Pregnancy

The physiologic effects of pregnancy and the puerperium include many hemodynamic changes, and the impact of these changes on AVM physiology remains controversial. A 50% increase in total body water persists throughout the pregnancy and declines 2 weeks postpartum. This hypervolemic state contributes to increased cardiac output, stroke volume, and heart rate. Systemic vascular resistance is decreased in pregnancy due to prostaglandin circulation, which lowers systemic arterial blood pressure. The decreased systemic vascular resistance in conjunction with increased venous compliance can result in increased venous stasis. Moreover, the inferior vena cava may be compressed by the gravid uterus, which leads to decreased blood return to the heart. Maternal cardiac output increases between 30% and 50% in pregnancy due to the increases in stroke volume and heart rate. Cardiac output rises 50% higher than baseline pregnancy values in the first stage of labor. Immediately after delivery, there is a 10%–20% rise in cardiac output, which then declines after the first postpartum day and returns to normal 2–4 weeks after delivery. Moreover, sex hormones such as estrogen may contribute to increased flow within the vessels that comprise an AVM, and this is hypothesized to increase the likelihood of AVM rupture. Estrogen may also lead to intimal hyperplasia in blood vessels. After uterine involution postdelivery, hormonal changes may also alter the structure of AVM vessels. These complex hemodynamic changes may each impact AVM biology at different stages of pregnancy, and understanding them can provide insight into the etiology and management of iAVM rupture in the pregnant patient.

Risk of iAVM Rupture and Hemorrhage in Pregnancy and the Puerperium

The lack of prospective studies limits assessment of the true risks associated with iAVM rupture to both the pregnant woman as well as the fetus. For this review, we searched the PubMed and Google Scholar databases from 1990 through 2021 for cerebrovascular AVM and pregnancy. The studies that we found are summarized below and on the next two pages.

Rupture appears likely to occur throughout pregnancy, not necessarily in the third trimester or during labor. In 36 cases of iAVM hemorrhage in pregnancy reported by Dias and Sekhar, only two cases occurred during childbirth. In a retrospective study of 451 female patients with AVMs who had 540 pregnancies at one institution in the United States, the incidence of hemorrhage was found to be 3.5% during the 52 weeks after the patient’s last menstrual period. This rate was not significantly different from the rate for similar nonpregnant populations.

In a cohort of 270 female patients with iAVMs at one institution in the United States, the annual hemorrhage rate in pregnant women was 5.7% compared to 1.3% in nonpregnant women, which represented a statistically significant increase in risk. This study suggests that patients with a known iAVM who become pregnant should be monitored carefully throughout pregnancy, particularly during the second and third trimesters and in the puerperium.

In a cohort-crossover study of three US states’ administrative claims data, each woman with an iAVM served as her own control, based on a 5-week period prior to a pregnancy of 40 weeks and 12 weeks of postpartum period. This study found that women with iAVMs had a 3.27-fold (relative risk; 95% confidence interval [CI], 1.67–6.43) increase in the risk of intracerebral hemorrhage during pregnancy and the puerperium compared with the nonpregnant period.

In a 2012 study of 54 female patients with iAVMs at one institution in the United States, there were a total of 62 pregnancies, and 5 hemorrhages occurred in 4 patients, yielding a hemorrhage rate of 8.1% per pregnancy.

A retrospective cohort study of 264 female patients with iAVMs at an institution in China analyzed hemorrhage during exposure periods (pregnancy and the puerperium) and nonexposure periods (the interval from birth until either AVM obliteration or last follow-up, after subtracting the exposure period) and found annual hemorrhage rates of 5.40% in exposure periods, 2.92% in nonexposure periods, and 3.82% in nonexposure periods of reproductive-age patients. The authors also performed a pooled data analysis, incorporating data from eight previously published studies and their own cohort. This analysis also showed a higher annual hemorrhage rate in exposure periods than in nonexposure periods (5.59% vs 2.52%; odds ratio, 3.19; 95% CI, 1.52–6.70).

A separate retrospective cohort study of 979 female patients with iAVMs in a different institution in China yielded contrasting findings. In this study, the odds ratio for AVM rupture during pregnancy and the puerperium, compared with the control period, was 0.71 (95% CI, 0.61–0.82). Thus the authors concluded that there was no increased risk of iAVM hemorrhage during pregnancy and the puerperium. Of the patients in this study who presented with intracranial hemorrhage due to AVM rupture during pregnancy, more than 91% presented in the second and third gestational trimesters, suggesting that later gestational age carries a greater risk of hemodynamic, coagulative, and vessel wall changes.

It is unclear whether the number of pregnancies a patient has had correlates with iAVM rupture risk. Most studies have not followed patients through multiple pregnancies, so the risk of AVM hemorrhage during subsequent pregnancies is unknown. However, there are some case reports of AVM rupture during a delivery after multiple normal pregnancies. One case report describes AVM hemorrhage in the patient’s fifth pregnancy, suggesting that risk persists even after multiple deliveries and that rapid diagnosis and treatment are required when intracranial hemorrhage is suspected in pregnant patients.

A 2019 systematic review analyzed data from three studies that provided a quantitative risk of first intracerebral hemorrhage due to iAVM in pregnancy. The authors extracted data on 47 cases across four cohorts and found that for these four cohorts, the annual risk of the first hemorrhage during pregnancy was 3.0% (95% CI, 1.7%–5.2%), 3.5% (95% CI, 2.4%–4.5%), 8.6% (95% CI, 1.8%–25%), and 30% (95% CI, 18%–49%); only the last cohort had a significant increase in risk compared to the risk in the nonpregnant period (relative risk 6.8; 95% CI, 3.6–13). The authors concluded that there is no conclusive evidence of an increased risk of the first hemorrhage due to iAVM during pregnancy and that a retrospective, multicenter case crossover study is “urgently required.”

There are rare reports in the literature of iAVM rupture occurring during labor and childbirth. One case report describes a patient at 41 weeks’ gestation who developed a sudden-onset severe headache during induction of labor for oligohydramnios. Her blood pressure when she complained of the headache had increased to 172/90 mm Hg. When her blood pressure remained elevated and the headache did not resolve, preeclampsia was suspected, and she was given a magnesium bolus, followed by magnesium infusion. A cesarean section was performed, resulting in the delivery of a healthy baby. During surgical closure of the cesarean section, the patient was noted to have a dilated pupil as well as poor uterine tone and oozing; she was intubated and hyperventilated. Intravenous administration of mannitol was initiated. Laboratory test results showed disseminated intravascular coagulation, and a CT scan of the patient’s head showed right temporal and frontal hemorrhage. Limited angiography ruled out a middle cerebral artery aneurysm, and an emergent craniotomy was performed for evacuation of intracerebral hemorrhage and resection of an AVM. The patient was in the intensive care unit (ICU) for 19 days, where she had little movement in her extremities, had limited speech (only a few words), and needed a gastrostomy tube for feedings. She was subsequently transferred to an inpatient rehabilitation setting for physical therapy and was described as “demonstrating slow, continued progress with aggressive physical therapy” when the case was reported. Another case report describes a patient who experienced convulsions during induction of labor and underwent brain MRI, which showed a cerebral AVM in the left frontal base without signs of intracranial hemorrhage or brain edema. The patient had a cesarean section and delivery of a healthy baby, with a plan to treat the AVM afterward.

Spontaneous regression of an iAVM is rare, and the occurrence during the puerperium has been described once in the literature. In this case, a pregnant patient presented with a ruptured AVM with a feeding artery aneurysm during the second trimester, and the AVM was clipped and hematoma removed. Serial angiography was performed 6 months and 1 year after delivery and showed dramatic regression of the AVM (described as “near-complete” at 1 year), suggesting that hormonal factors during pregnancy and the immediate postpartum period affect the growth and rupture of an AVM.

Patients who present with hemorrhage due to iAVM rupture during pregnancy may have an elevated risk of rebleeding during the same pregnancy compared to the risk of early rebleeding in nonpregnant patients. One study included 27 pregnant women with AVM rupture who did not have immediate resection of the AVM, and 7 of these women had hemorrhage recurrence before or immediately after delivery, yielding a rebleeding rate of 26%. The Stroke Council of the American Stroke Association acknowledges that this rebleeding rate is higher than the 6% expected rebleeding rate in the first year after a hemorrhage in nonpregnant patients. Although the evidence is only level V, these findings suggest there may be a benefit to early therapy in some pregnant patients presenting with AVM rupture.

In a literature review published in 2016, Lv et al. analyzed data from 65 cases of AVM in pregnancy and found that AVM hemorrhage presentation was significantly associated with a poor maternal outcome (modified Rankin Scale score ≥2) but was not significantly associated with risk to the fetus. Additionally, gestational age at the time of AVM hemorrhage was not significantly associated with either poor maternal outcome or fetal risk.

A retrospective review of the University of California San Francisco Brain Arteriovenous Malformation Project database identified 16 cases in which patients with AVMs became pregnant. In three of the cases, the pregnancy did not continue past the first trimester. In the remaining 13 cases, 10 patients underwent emergent AVM treatment before delivery and 3 deferred treatment until after delivery. Eleven (85%) of the 13 patients had good maternal outcomes (modified Rankin Scale score ≥2), including 8 of the 10 patients whose AVMs were treated before delivery. There were no reports of postnatal cognitive or developmental delays in infants or toddlers at 2-year follow-up.