Vasa previa—velamentous fetal vessels over or near the cervix—occurs in 1 in 2,500 pregnancies. The criterion used for ultrasonographic diagnosis in most studies is velamentous vessels within 2 cm of the cervix. Risk factors include multiple gestation, in vitro fertilization (IVF), previa early in pregnancy, and bilobed or succenturiate placenta.1–3 More than half of undiagnosed cases result in fetal or neonatal death.4 Prenatal diagnosis was first reported in 1987,5 and there have been numerous reported cases, including eight single-institution series of five or more patients1,3,6–11 and collaborative series of 47 and 61 patients.4,12 Sensitivity and specificity of ultrasonographic diagnosis in a recent review are greater than 90% and nearly 100%, respectively.13 Prior series have not shown an association of vasa previa with fetal anatomic abnormalities.
Management recommendations are still evolving. Publications draw from expert opinion, retrospective studies, and decision analysis.14–18 Most authorities recommend corticosteroids at 28–32 weeks of gestation, cervical length surveillance, and consideration of inpatient management from approximately 30 weeks of gestation. Delivery timing is controversial; recent articles suggest delivery at 34–35 weeks of gestation,15–17 timing based on cervical length but before 36 weeks of gestation,18 and between 34 and 37 weeks of gestation.14 Twin gestation is a risk factor for both premature birth and vasa previa, but there are limited data regarding outcomes of twins with vasa previa and management recommendations have not been defined for twins compared with singletons. We conducted a retrospective study of our institutional experience with maternal and neonatal outcomes in prenatally diagnosed cases of vasa previa, with consistent placental pathology, over 12 years to expand the evidence base regarding vasa previa.
MATERIALS AND METHODS
After approval by the Sharp health care institutional review board, we identified possible cases of prenatally diagnosed vasa previa with delivery at Sharp Mary Birch Hospital for Women and Newborns from July 2003 to June 2015 using 1) a medical records search using International Classification of Diseases, 9th Revision codes 633.50, 633.51, and 633.53; and 2) perinatal case lists. For potential patients thus identified, we reviewed the admission, delivery, discharge summaries, and pathology reports. Patients were evaluated for inclusion if there was prenatal ultrasonographic diagnosis of vasa previa. Patients in whom pathologic examination of the placenta was not conducted or in which pathology was nondiagnostic or inconclusive as a result of placental disruption were excluded from analysis. For patients with confirmed diagnosis at delivery and consistent pathologic examination, detailed maternal and neonatal chart review was conducted. False-positive cases, in which the diagnosis of vasa previa was suspected by ultrasonography but not confirmed by the pathologist's findings, were not included in the final data set but are presented separately in the “Results” section.
Vasa previa is defined as velamentous vessels of the cord within 2 cm of the cervix. Vasa previa was classified as type 1 when there was a velamentous insertion of the umbilical cord into the placenta with vessels overlying the cervix, type 2 when there was a multilobed or succenturiate placenta with fetal vessels connecting two lobes overlying the cervix, and type 3 when there were velamentous vessels overlying the cervix without either a velamentous cord insertion or succinturate lobes.1,2 Placenta previa was defined as placental edge within 2 cm of the cervix on the last ultrasonogram before delivery. Patients in whom vasa previa was suspected based on prenatal ultrasonography and confirmed at delivery formed the study population. Patients were excluded from analysis if there was a concomitant finding of placenta previa. One patient, in whom type 2 vasa previa had been treated by laser ablation of the vessels over the cervix (and thus was not a vasa previa at the time of delivery), was excluded.
Sharp Mary Birch Hospital for Women and Newborns is a tertiary care center serving San Diego, Imperial, and southern Riverside counties. This is a private nonprofit hospital. Ultrasonographic screening is by radiologists, specialist obstetrician–gynecologists, and maternal-fetal medicine subspecialists. Women with suspected vasa previa are generally referred for maternal-fetal medicine confirmation and consultation. Prenatal diagnosis of vasa previa was made using real-time and color Doppler imaging confirmed by transvaginal ultrasonography with periodic transvaginal cervical length assessments at the discretion of the maternal-fetal medicine subspecialist.
Ultrasound systems used included Ultramark 9, Acuson Sequoia, Siemens Antares, and Voluson 730 and E8. Treatment including antepartum hospitalization, use of tocolysis, administration of steroids for acceleration of lung maturity, and timing of delivery were at the discretion of the attending obstetrician–gynecologist and consulting maternal-fetal medicine specialist. General recommendations included serial ultrasound scans, elective hospitalization by 30–32 weeks of gestation, and hospitalization nonelectively for cervical shortening, bleeding, or contractions. From 2004 to 2011 delivery was recommended at 35–36 weeks of gestation based on a large multicenter study.4 After publication of a decision analysis19 in 2011 suggesting that the optimal delivery timing was at 34–35 weeks of gestation, one of the maternal-fetal medicine groups recommended elective delivery 34–35 weeks of gestation with exceptions based on the presence of major anomalies.
We reviewed maternal medical records and abstracted data including maternal demographics, conception type, number of fetuses, presence or absence of placenta previa after 16 weeks of gestation, gestational age at which the diagnosis of vasa previa was made, prenatal determination of type of vasa previa, and fetal anomalies diagnosed by ultrasonography. Final determination of vasa previa type was based on findings at delivery and placental pathology description. Data regarding course of pregnancy included gestational age and indications for antepartum admission, use of tocolysis and administration of corticosteroids, gestational age and indication for delivery, antepartum and postpartum hospital stays, and maternal complications. Delivery was categorized as elective if the only reason for delivery was the presence of a vasa previa and attainment of gestational age goal and otherwise as nonelective. Placental findings were taken from the operative note and from the results of placental pathology. Hospital records were also reviewed for near misses, which were defined as ruptured membranes prompting delivery, a dilated cervix at the time of delivery, or bleeding resulting in need for urgent transfusion of the newborn.
Information collected from neonatal medical records included gestational age at delivery, birth weight and Apgar scores, need for urgent neonatal transfusion (within 12 hours of birth), duration of hospital stay, and perinatal mortality. We recorded complications of prematurity including respiratory distress syndrome, intraventricular hemorrhage, and necrotizing enterocolitis and major and minor anomalies from diagnoses in the admission and discharge summaries.
Data were incomplete for some patients regarding gestational age at which diagnosis of vasa previa was made (n=3), administration of corticosteroids (n=2), and major and minor anomalies noted in the admission or discharge summary (n=4).
Data were entered into a computerized database. Statistical calculations were performed using SPSS 22 and Graphpad. We report categorical variables as mean, standard deviation, and range; statistical comparisons were made using the Student t test, χ2, Fisher exact test, and Mann-Whitney U test. Statistical significance was defined as P<.05.
There were 96 pregnancies of which 77 were singletons and 19 were twins (12 monochorionic and seven dichorionic) with prenatally diagnosed vasa previa and consistent placental pathology delivered at Sharp Mary Birch Hospital for Women and Newborns during the study period.
There were four possible false-positive diagnoses. In the first, placenta previa was seen at 26 weeks of gestation and vasa previa without placenta previa was suspected by ultrasonography at 31.5 weeks of gestation. Delivery was at 34 weeks of gestation as a result of vaginal bleeding. Placental pathology showed a normal cord insertion and no accessory lobes. The newborn weighed 2,805 g and was discharged on day 9. In the second, vasa previa for twin A of dichorionic twins was suspected at 28 weeks of gestation, not visualized on transvaginal ultrasonogram at 30 weeks of gestation, and a final ultrasonogram at 31.5 weeks of gestation was inconclusive regarding vasa previa compared with marginal sinus. Cesarean delivery was performed for preterm labor at 34.5 weeks of gestation. Birth weights were 2,470 and 1,810 g, and hospital stays were 16 and 20 days. The placentas had normal cord insertions and no accessory lobes. Ultraonographic images were not available for review for these two patients.
There were two additional patients in whom digital images were reviewed and consistent with the diagnosis of vasa previa. In the first, on multiple transvaginal scans starting at 27 weeks of gestation, the umbilical cord appeared to form within the membranes at the lower placental edge within 2 cm of the cervix. Digital images from one of the scans were reviewed by the authors (V.A.C., W.S., L.C.) who confirmed this finding. Prenatal ultrasonography also showed hydrocephalus, porencephaly, and a grade 4 intracranial hemorrhage. Elective cesarean delivery was performed at 35 weeks of gestation. The neonate weighed 2,950 g and had a 9-day hospital stay. At pathology, the cord insertion was described as marginal just at the placental edge, but not velamentous. This may represent a semantic distinction rather than a false-positive. In the second, a 2-mm velamentous vessel overlying the cervix with a venous flow signature within was suspected at 23 weeks of gestation and was demonstrated on subsequent transabdominal and transvaginal ultrasonograms. The neonate was electively delivered at 35 weeks of gestation, weighed 2,460 g and required 9 days in the hospital. Image review by the authors (V.A.C., W.S., L.C.) confirmed the ultrasonographic findings. The vessel was not tagged at delivery and was not identified on placental pathologic examination. This may represent a false-positive ultrasonographic diagnosis or a false-negative pathologic examination. All four cases are excluded from the final data set.
Total delivery volume at Sharp Mary Birch Hospital for Women and Newborns in the study period was 100,481 pregnancies, 97,907 singleton and 2,574 twin. The rates of vasa previa were 1 of 1,272 for singleton pregnancies and 1 of 135 for twin pregnancies (P<.001). Six of the 19 twin pregnancies (28.5%, 95% confidence interval [CI] 11–52%) and 16 of the 77 singletons (20.8%, 95% CI 12–31%) were conceived by IVF. Maternal demographics, gestational age at diagnosis, and antepartum hospitalization are summarized in Table 1. Gestational age at diagnosis of vasa previa without placenta previa for singletons was at 22.9±5.2 weeks of gestation with the earliest diagnosis at 13.2 weeks of gestation and the latest at 38.4 weeks of gestation. For twins, corresponding figures were 22.6±4.7 weeks of gestation with a range of 15–33 weeks of gestation. Among the 96 pregnancies, 70 (72.9%) had type 1 vasa previa; type 1 vasa previa was present in 15 of 19 twins (78.9%) and 55 of 77 singletons (71.4%). Type 3 vasa previa was diagnosed in one twin pregnancy and 25 were type 2.
The major risk factors for vasa previa in prior studies1–3,12 are IVF pregnancy, multiple gestation, placenta with succenturiate lobe or lobes, and placenta previa diagnosed earlier in the pregnancy, but resolved before delivery There were 19 cases—24.7% of the 77 singletons—with none of these risk factors; all had type 1 vasa previa.
Cases per year of prenatally diagnosed vasa previa ranged from 6 to 14. The median number of cases per year from 2004 to 2009 was 6, and the median from 2010 to 2014 was 11 (P<.05, Mann-Whitney U test).
Ninety-two of the 96 women with vasa previa were hospitalized antenatally (94.8%). Among the 77 singleton pregnancies, 44 (57.1%) were electively hospitalized at 31.1±2.3 weeks of gestation with estimated gestational age at delivery 34.7±1.2 weeks. Twenty-nine (37.77%) women with singletons had nonelective hospitalization at 31.9±3.2 weeks of gestation with delivery at 34.4±1.9 weeks of gestation. Four (5.2%) were not hospitalized antepartum and were admitted for elective cesarean delivery at 36–38.4 weeks of gestation, including one patient diagnosed at 38.4 weeks of gestation. All of the women with twin pregnancies had antepartum hospitalization. Nine (47%) had elective hospitalization at 27.7±2.0 weeks of gestation and delivered at 33.1±1.3 weeks of gestation. Ten (53%) had indicated hospitalization at gestational age 28.2±4.8 weeks and delivered at 32.5±2.7 weeks of gestation.
Tocolysis was used for 52 of the 77 singletons 67.5% with first treatment at a gestational age of 30.6±3.2 weeks. Eighteen of the 19 twin pregnancies (95%) received tocolysis, starting at 26.0±3.3 weeks of gestation. Tocolytics used included magnesium sulfate, nifedipine, indomethacin, and terbutaline. We did not collect data regarding the rates of tocolysis for acute control of contractions as opposed to administration of maintenance tocolysis. Betamethasone was given in 65 of 77 (84.4%) of women with singletons and 18 of 19 (95%) with twins. Gestational age at administration of betamethasone was 30.5±2.4 weeks (range 23.5–34 weeks of gestation) for singletons and 26.8±2.3 weeks of gestation (range 23.2–32.2 weeks of gestation) for twins. We did not collect data regarding the use of a second course of betamethasone. All of the patients delivered at or before 34 weeks of gestation received betamethasone before delivery.
Delivery timing and indications for singleton and twin pregnancies are summarized in Table 2. Elective delivery occurred in 48 women with singletons. Delivery was indicated in 29 women in this group: 21 for contractions or labor, bleeding in four (including one with ruptured membranes and a nonreassuring tracing), nonreassuring tracing in two, asymptomatic cervical shortening in one, and preeclampsia in one. One of the singleton neonates, delivered for bleeding and ruptured membranes with a nonreassuring tracing, required urgent neonatal transfusion. Among the 19 women with twins, delivery timing was elective in six and indicated in 13 resulting from contractions or labor, including two with a dilated cervix (1.5 and 3 cm).
The rate of indicated delivery was significantly higher in twins than in singletons: 13 of 19 (68%) compared with 29 of 77 (38%) (P<.05). There were near misses in 2 of 19 twin deliveries (10.5%, 95% CI 1.3–33.1%), which was not statistically different from that of singletons, which was 1 of 77 (1.3%, 95% CI 0–7.0%) (P>.10). Nonelective deliveries were significantly more frequent in twins than singletons at each gestational age examined: before 32 weeks of gestation (5/77 [6.5%] in singletons compared with 5/19 [26.3%] twins, P<.05); before 33 weeks of gestation (9/77 [11.7%] singletons compared with 8/19 [42.1%] twins, P<.01), and before 34 weeks of gestation (9/77 [11.7%] singletons compared with 11/19 [58%] of twins, P<.001).
We compared elective delivery timing for singletons treated up to 2011 and from 2012 onward. Among 48 women who were carrying singletons, who had vasa previa diagnosed before 34 weeks of gestation, and who were electively delivered, cesarean deliveries occurred before 35 weeks of gestation in 4 of 20 (20%) treated in 2011 or before and in 10 of 27 (37%) treated in 2012 or after. This difference did not reach statistical significance.
There were 13 pregnancies delivered after 36 weeks of gestation. Three had late diagnoses at 36.2–38.4 weeks of gestation and were promptly delivered; one was treated in 2004, before our recommendation for delivery by 35–36 weeks of gestation. Among the remaining nine, there were three clinically stable patients who were carrying fetuses with serious congenital heart defects (transposition, hypoplastic left heart, and a dichorionic twin set with A having mitral and aortic stenosis and coarctation). After risk and benefits were reviewed with pediatric cardiology, the decision was made for close inpatient monitoring and delivery at 37 weeks of gestation. In six, including a set of monochorionic twins delivered at 36.4 weeks of gestation, there was no reason given for the late delivery.
There was one serious maternal complication, a deep venous thrombosis, which occurred in a 42 year old who had been hospitalized for 25 days as a result of cervical shortening. She was treated with anticoagulation and a removable inferior vena cava filter. She was delivered at 34 weeks of gestation with an uneventful postoperative course.
Neonatal data and length of stay are summarized in Table 3. All of the neonates were live-born. For singletons, the gestational age at delivery was 34.7±1.6 weeks with mean birth weight 2,367±447 g. For twins, gestational age was 32.8±2.2 weeks. Although we could not be certain from operative records that twin A was always the neonate with vasa previa, we separately calculated birth weights, which were similar: 1,860±525 g for the first twin and 1,955±492 g for the second twin.
Excluding neonates with anomalies, hospital stays were 15.8±16.5 days with a range of 2–82 days. For nonanomalous singletons, the hospital stays were 11.6±10.8 days with a range of 2–61 days. Nonanomalous twins had an average stay of 26.2±22.5 days with a range of 3–82 days (P<.001 singleton compared with twin).
Gestational age, birth weight, and length of stay are tabulated by gestational age for neonates without anomalies in Table 4.
Maternal demographics and delivery outcomes were compared for spontaneous and IVF pregnancies. Women with singleton spontaneous pregnancies were younger than those with singleton IVF pregnancies—32.4±5.9 years compared with 36.2±5.0 years, but this difference did not reach statistical significance (P=.06). Gestational age at delivery and birth weight for spontaneous compared with IVF singletons were similar at 34.7±1.7 weeks of gestation compared with 34.7±1.1 weeks of gestation and 2,374±482 g compared with 2,327±285 g.
Comparisons for the 13 spontaneous and six IVF twin pairs were limited by sample size. Maternal age was 29.2±6.9 years compared with 33.3±5.9 years. Gestational age at delivery were 33.0±1.8 and 32.4±2.8 weeks. Birth weights of the first and second twins were 1,881±429 g compared with 1,815±685 g for twin A and 2,034±441 g compared with 1,786±550 g for twin B, respectively. None of these differences reached statistical significance but the study was underpowered for these comparisons.
Fetal anomalies are summarized in Table 5. Major anomalies were present in six neonates—3 of 38 twins (7.9%, 95% CI 1.7–21%) and 3 of 77 singletons (3.9%, 95% CI 0.8–11.0%). All of these were diagnosed in utero. Among dichorionic twins, there was one case of transposition of the great arteries and one case of mitral and aortic stenosis with coarctation of the aorta, both in IVF pregnancies. Among monochorionic twins, there was one with gastroschisis. Among IVF singletons, there was one neonate each of a large neck mass with a differential diagnosis at discharge of hemangioma compared with malignancy. Among spontaneous singletons, there was one case of transposition of the great arteries and one neonate with a hypoplastic left heart and atrial septal defect.
Minor anomalies diagnosed in utero included three neonates with pyelectasis, all in spontaneous singletons. Additional anomalies diagnosed after birth in IVF singletons were one neonate with mild branch pulmonic stenosis and one neonate with mild coarctation of the aorta. One IVF twin had mild ventriculomegaly. In spontaneous singletons, there were two neonates with secundum atrial septal defects and one neonate with scoliosis with hemivertebra. Among non-IVF monochorionic twins, there was one small muscular ventricular septal defect, which resolved before hospital discharge, and one neonate with a Dandy-Walker variant.
Three (10.7%) of the 28 IVF newborns had major anomalies compared with 3 (3.5%) of 87 non-IVF newborns. Four (14.2%) of the 28 IVF neonates had cardiac anomalies compared with 5 (5.7%) of 87 of the non-IVF newborns. Six (21.4%) of the IVF newborns compared with 11 (12.6%) of the 87 non-IVF neonates had any anomaly. None of these differences reached statistical significance.
One singleton neonate required urgent transfusion after delivery. The mother had been electively hospitalized at 27 weeks of gestation. At 32.5 weeks of gestation, she had ruptured membranes with bloody amniotic fluid and fetal distress. An emergency cesarean delivery was performed. The neonate had Apgar scores of 7 and 8 but a hematocrit of 29.9%. The neonate was urgently transfused and did well. Hospital stay was 22 days.
Neonatal respiratory distress syndrome was diagnosed in 44 of 77 (57.1%) of singletons and 25 of 38 (65.7%) of twins. There was one case each of necrotizing enterocolitis, pneumothorax requiring thoracotomy, and focal basal ganglia infarct.
There were no perinatal deaths resulting from ruptured vasa previa (95% CI 0–3.1%). There was one neonatal death (1 of 115 neonates; 0.9%, CI 0–4.8%), a child with a complex heart defect including hypoplastic left heart and atrial septal defect. Delivery was uncomplicated, but the infant died on day 28 after cardiac surgery.
Our data support the conclusion that outcomes with vasa previa can be excellent. Among our 96 pregnancies and 115 neonates, the only neonatal death was the result of congenital heart disease. There was 100% survival among 129 patients in five recent series.3,8–10,12 Even with a prenatal diagnosis, catastrophes occur. In two series of greater than 50 cases, 5 of 117 pregnancies had deaths from ruptured vasa previa.4,11
Elective preterm delivery is recommended in cases of vasa previa to avoid the catastrophic potential consequence of ruptured vasa previa. Determination of the optimal timing for delivery balances risks of prematurity against those of labor or rupture of membranes. Robinson and Grobman,17 in their decision analysis, estimated spontaneous preterm birth rates of 1.5% per week and 4% per week at 32–34 and 34–36 weeks of gestation and suggested delivery at 34–35 weeks of gestation. They did not analyze separately for twins. Singleton indicated delivery rates in our series are higher than in their decision analysis: 4 of 72 from 32 to 34 weeks of gestation or 2.8% per week and 12% and 17% at 34–34.9 and 35–35.9 weeks of gestation (Table 2). As indicated, delivery resulting from contractions and bleeding increased at 34 weeks of gestation in our population; we concur with recommendations by Robinson and Grobman17 and Swank et al12 to deliver singletons by 34–35 weeks of gestation.
Four prior series had five or more five sets of twins.4,10–12 Only Bronsteen et al11 separated outcomes. Median delivery gestational ages for seven twin and 49 singleton pregnancies were 32 compared with 35 weeks with emergent delivery rates of 28.6% compared with 4.1%. Two of three perinatal deaths as a result of vasa previa were in twins. In our series, twins delivered 2 weeks earlier than singletons. Indicated deliveries for twins were 5.3–11.8% per week at 28–32 weeks of gestation, 21% at 32–33 weeks of gestation, and 30% at 33–34 weeks of gestation. The small number of twins in our study precludes determination of the ideal time to delivery twins with vasa previa, but it appears that risks for indicated delivery increase numerically at 32 weeks of gestation.
Elective hospitalization is controversial.8,9,12,16 For example, Swank et al12 argues in favor and Hasegawa et al9 against. The Society for Maternal-Fetal Medicine does not find adequate data to support a recommendation.14 Rupture of a fetal vessel within the membranes with vasa previa may lead to rapid blood loss from the fetus, resulting in intrauterine or neonatal demise. Inpatient management affords the potential for a rapid move to the operating room for emergency cesarean delivery. In our series, there was one case in which elective hospitalization was likely life-saving—rupture of membranes, bleeding, and fetal distress prompting emergency cesarean delivery.
Although catastrophic outcomes of vasa previa may occur in the hospitalized patient, it is unlikely that a trial of inpatient compared with outpatient management will soon be done, and it seems unlikely that outpatient management would be safer for the fetus. We offer all patients with singleton pregnancies and vasa previa elective hospitalization at 30–32 weeks of gestation. Extending the reasoning of Swank et al,12 based on the high rates of indicated deliveries even at 28 weeks of gestation in twins, and pending larger studies of the rates of need for intervention in twins with vasa previa, we suggest that twins be offered hospitalization earlier—by 26–28 weeks of gestation.
Most series4,8,10–12 do not give detailed neonatal outcomes. Two3,9 report anomalies; among 30 cases of vasa previa, there one case each of ventricular septal defect and duodenal atresia. In our prior report of 10 cases, there was one major cardiac anomaly.1 Summing these series, 3 (7.5%) of 40 neonates had anomalies. We found a high rate of major fetal abnormalities, particularly cardiac. Of 115 neonates, six (5.2%) had major anomalies and 17 (14.7%) had any anomaly. Four (3.5%) had major heart defects and nine (7.8%) had any cardiac anomaly. Monochorionic twinning is reported to have a markedly increased risk for anomalies including congenital heart defects,19 but only one of the cardiac anomalies in our series was in a monochorionic twin set. In most published studies, IVF carries a slightly increased risk for cardiac anomalies20; however, IVF pregnancies in our study had a 14.2% risk for congenital heart defects. Whether the high rates of anomalies in our study represent an independent association with vasa previa or other factors will require larger data sets for resolution.
The current study has several strengths. We have presented the maternal and perinatal outcomes of a relatively large number of patients with confirmed vasa previa managed at a single institution over 12 years. Because of the single site, there was access to maternal and neonatal charts. In addition, there is a larger number of twins than in prior reports. The weaknesses of our study are that the report is retrospective and, as such, some data are missing and all of it is based on medical record charting As well, management changed during the study and differed between managing physician groups, potentially introducing bias. Although the number of twins reported is larger than in other reports, it is still too small to allow for meaningful analysis.
In our series, there were 19 cases of vasa previa in patients without risk factors. The American Institute of Ultrasound in Medicine recommends that “the placental cord insertion site should be documented when technically possible”21 and the Society for Maternal-Fetal Medicine recommends transvaginal ultrasonography in cases of resolved previa.14 These should result in detection of most cases of vasa previa. If the cord insertion is not visualized, a color Doppler evaluation over the cervix may improve ascertainment.
There were four patients in this series with vasa previa not confirmed by pathology. Two and possibly three were false-positive diagnoses. Given the importance of the diagnosis, vasa previa should be confirmed on serial transvaginal ultrasonograms, and there should be close communication between the obstetrician and pathologist for placental examination.
The number of prenatally diagnosed vasa previa cases in this series rose from a median of six per year from 2004 to 2009 to a median of 11 per year from 2010 to 2014. There was not a parallel rise in the delivery volume at Sharp Mary Birch Hospital for Women and Newborns. We cannot be certain of the reason for the increase, but we suspect that better awareness of the diagnosis, improving image quality of ultrasound machines, and an increasing volume of prenatal diagnostic studies performed by maternal-fetal medicine subspecialists may have contributed.
Nearly four decades ago, Pent predicted: “a moderately active obstetrics service can expect to have one perinatal death per year due to vasa previa.”22 We speculated in 2001 that a busy service might prevent one death per year.1 In one series comparing outcomes between prenatally diagnosed and undiagnosed vasa previa,4 survival rates were 97% and 44%. If the mortality rate is indeed halved by prenatal diagnosis, we may have prevented nearly 50 perinatal deaths over the past 12 years.
Among the approximately 4 million yearly births in the United States, there are an estimated 1,600 cases of vasa previa. To improve outcomes, vasa previa must be diagnosed and preventive strategies used to attempt to avoid rupture of the vasa previa vessels before delivery and balance this against degree of prematurity. Unfortunately, management recommendations will be based on observational rather than randomized trial data as a result of the rarity of vasa previa.
Based on the results of our study, we recommend targeted ultrasonography and fetal echocardiography after a diagnosis of vasa previa. Elective hospitalization should be offered for proximity to an operating room. We concur with others' recommendations to deliver singletons with vasa previa by 34–35 weeks of gestation. Consideration of earlier delivery in twins is appropriate and should be individualized for the patient. Prophylactic steroids proximal to planned early delivery is appropriate, including a single “rescue” course if indicated.23 If steroids were not given before 34 weeks, a single course should be given.24 As a result of the infrequent nature of this disorder, collaborative research to accumulate outcomes for twins in particular is needed. Future studies should include neonatal follow-up.