- Email: [email protected]
Periviable birth at 20 to 26 weeks of gestation: Proximate causes, previous obstetric history and recurrence risk
American Journal of Obstetrics and Gynecology (2005) 193, 1175–80
www.ajog.org
Periviable birth at 20 to 26 weeks of gestation: Proximate causes, previous obstetric history and recurrence risk Brian Mercer, MD,a Cynthia Milluzzi, BFA, BSN,a Marc Collin, MDb Departments of Reproductive Biologya and Pediatrics,b MetroHealth Medical Center at Case Western Reserve University, Cleveland, OH Received for publication March 1, 2005; revised May 2, 2005; accepted May 9, 2005
KEY WORDS Preterm birth Periviable birth Prematurity Prediction
Objective: Early preterm birth at 20 to 26 weeks of gestation (periviable birth) carries extreme risks of infant death and morbidities. Prevention of periviable birth could improve infant outcomes significantly. We sought to characterize the causes of periviable birth and to determine whether periviable birth can be predicted by previous pregnancy outcome. Study design: We evaluated 104,921 pregnancies (1974-2004) and assessed the frequency and causes of periviable birth. Women who were delivered of both their first and second pregnancies at O20 weeks of gestation at our institution were identified. Predictive values of the first pregnancy outcomes for second pregnancy outcomes were determined. Results: Periviable birth complicated 1981 deliveries (1.9%). Seventy-nine percent of the women with periviable births had no history of periviable births; 44% of the women had no previous deliveries, and 35% of the women had previous term deliveries only. Causes of periviable birth were labor (36%), premature rupture of membranes (34%), bleeding (10%), and preeclampsia (4%). Four percent of the gestations were multiple gestations. Among 7970 pregnancies at O20 weeks of gestation, periviable birth in the first pregnancy was associated with preterm birth and periviable birth in the second pregnancy (35.6%, 6.9%; relative risk, 3.3 and 8.6; P ! .0001). Periviable birth and preterm birth in the first pregnancy were insensitive for periviable birth in the second pregnancy (8.8%, 36.8%, respectively). Conclusion: Although periviable birth is associated with subsequent periviable birth and preterm birth, preterm birth and periviable birth are insensitive markers for recurrences in the next pregnancy. Early pregnancy or preconceptional markers for prediction of periviable birth are needed. Ó 2005 Mosby, Inc. All rights reserved.
Despite efforts directed towards prevention of preterm birth, the incidence of prematurity has increased by one-third from 9% to 12% over the past 2 decades and
Presented at the 25th Annual Meeting of the Society for Maternal Fetal Medicine, February 7-12, 2005, Reno, Nevada. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2005 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2005.05.040
remains twice as common among African American births as white births.1-3 Most perinatal death results from complications of prematurity. Long-term sequelae that include neurologic handicaps, blindness, deafness, and chronic respiratory disease are linked directly to preterm birth, particularly before 30 weeks of gestation.4-6 Very low birth weight infants (!1500 g) account for 64% of all neonatal deaths.7
1176 Preterm birth that occurs near the limit of potential viability carries profound implications for the infant, family, and society. Survival is not anticipated with delivery before 23 weeks. Thereafter, survival increases progressively and steeply with each week that is gained between 23 and 30 weeks of gestation.8,9 With extreme prematurity, most survivors experience significant acute morbidities and are at risk for long-term sequelae.5,9,10 However, acute and subacute perinatal morbidities decrease with relatively small increases in gestational age at delivery from 23 weeks of gestation.7-11 Women suffering a preterm birth are at increased risk for recurrence in subsequent gestations.12 This risk increases with an increasing number of and decreasing gestational age at previous preterm births.12-15 Further, the risk of early preterm birth also increases with decreasing gestational age of the earliest previous preterm birth.14 However, studies regarding early preterm birth have disagreed as to the prevalence of preterm labor, premature rupture of the membranes, and other causes.5,16-18 The prevention of early preterm birth offers the greatest potential to reduce infant morbidity and mortality rates. Acute interventions to prevent preterm birth are not highly effective once symptoms occur. Because of this, we undertook this study to characterize the causes of admission before early preterm birth and to determine the ability of previous outcome to predict subsequent pregnancy early preterm birth.
Material and methods This retrospective cohort study was approved by the Institutional Review Board of MetroHealth Medical Center, Case Western Reserve University. We evaluated data that were collected between December 1974 and June 2004 by chart review at discharge and stored in our computer-based perinatal database. We evaluated all pregnancies that delivered a live- or stillborn infant at R20 weeks of gestation. Those infants with birth weights !150 g were considered implausible and were excluded.19 Our analysis was restricted to elements that had been collected continuously over the 30-year existence of this database and included gravidity, parity, admission indication, admission/delivery date and time, admission cervical dilation, delivery gestation (caregivers’ best obstetric estimate), and birth weight. Admission indications included labor, preterm rupture of membranes (PROM), incompetent cervix, vaginal bleeding, preeclampsia, fetal abnormalities, fetal death or growth restriction, and ‘‘other complications.’’ Twin/multifetal pregnancies were included once for each pregnancy and were considered separately. We defined periviable birth as a delivery that occurred between 20 weeks 0 days of gestation and 26
Mercer, Milluzzi, and Collin
Figure Previous obstetric outcome among women who were delivered at 20 weeks to 26 weeks 6 days of gestation in the current pregnancy.
weeks 6 days of gestation, with a birth weight between 150 and 1499 g.19 All pregnancies that met the inclusion criteria were evaluated to determine the frequency of periviable birth, and admission indications were determined for this subgroup. We then selected, from the aforementioned cohort, only those women who were delivered of both their first and second pregnancy at R20 weeks of gestation at our institution (G2P2s). Women with any miscarriages or therapeutic abortions and women with an intervening pregnancy between the first and second pregnancy were excluded, as were women with twin/ multifetal gestations in either pregnancy. We determined the likelihood of preterm and periviable delivery in the second pregnancy on the basis of the presence or absence of a similar history in the first pregnancy. We subsequently evaluated the time from admission to delivery for women with a periviable birth to determine the opportunity for intervention to prevent delivery or maximize fetal/neonatal outcomes. This information was calculated from the times and dates of admission and delivery (date, hour, minutes), which was documented in the database between January 1994 and June 2004. Descriptive statistical analyses were performed with Statview software (version 5.0.1; SAS Institute Inc, Cary, NC). Ninety-five percent confidence intervals for point estimates were determined by the method of Wilson.20 Fisher’s exact test was used to compare outcome frequencies. A probability value of !.05 was considered significant.
Results Of 104,921 pregnancies that were delivered between December 1974 and June 2004, periviable birth at 20
Mercer, Milluzzi, and Collin Table I
1177
Point estimates for various causes of early preterm birth in 5 studies
Inclusion criteria
Fliegner et al12 (1987) 22-28 wk
Tucker et al13 (1991) 20-29 wk
Gray et al14 (1997) 24-29 wk
Bottoms et al5 (1999) %1000 g
Mercer et al (2005) 20-26 wk
Sample size (n) Premature labor (n) Premature rupture of the membranes (n) Antepartum hemorrhage (n) Fetal death (n) Hypertension/preeclampsia (n) Multiple pregnancy (n) Congenital malformations (n) Incompetent cervix (n) Fetal growth restriction (n) Chorioamnionitis (n) Fetal distress (n) Other or unexplained (n)
d* 18 12 13 d* 7 14 11 13 d* d* d* 12
354 23 33 4 13 2 d* 2 d* 4 2 2 17
189 14 34 20 d* 14 d* d* d* d* 8 d* 10
713 42 22 4 d* 12 d* d*
1981 36.3 (34.2-28.4)y 34.1 (32.0-36.2)y 10.4 (9.1-11.8)y 6.2 (5.2-7.4)y 4.0 (3.2-4.9)y 4.0 (3.2-4.9)y 2.4 (1.8-3.2)y 0.5 (0.3-0.9)y 0.2 (0.08-0.5)y d* d* 2.0 (1.5-2.7)y
2 d* 9 9
* Not evaluated. y Data given as percent (95% CI).
Table II Predictive value of periviable birth at 20 to 26 weeks of gestation in the first pregnancy for preterm birth at 20 to 26, 20 to 30, 20 to 34, and 20 to 36 weeks of gestation in the second pregnancy Gestation at delivery in second pregnancy Variable
20-26 Wk
20-30 Wk
20-34 Wk
30-36 Wk
Periviable birth in first pregnancy (%) No previous periviable birth (%) Relative risk of periviable birth in the second pregnancy P value Sensitivity of periviable birth for subsequent preterm birth (%)
6.9 0.8 8.6 !.0001 8.8
11.5 1.8 6.4 !.0001 6.8
19.5 4.9 4.0 !.0001 4.5
35.6 10.9 3.3 !.0001 3.5
weeks to 26 weeks 6 days of gestation complicated 1981 deliveries (1.9%). One third of these deliveries (33.1%) occurred before the limit of viability (20 weeks-22 weeks 6 days of gestation). The incidences of preterm birth at 20 to 30, 20 to 34, and 20 to 36 weeks of gestation were 4.0%, 9.3%, and 16.2%, respectively. Women with periviable births were much more likely to have had R2 previous spontaneous or induced abortions than those without a periviable birth (23.8% vs 7.1%; P ! .0001). Although more likely to have had a previous preterm birth than women without such a history (21.1% vs 15.6%; P ! .0001), only 1in 5 women with a periviable birth had a history of R1 previous preterm births. Among these, 52.6% of the women had R1 previous term deliveries in addition to the previous preterm birth. Forty-four percent of the women had no previous deliveries (25.5% were in their first pregnancy, and 18.9% had R1 previous miscarriages and/or therapeutic abortions but no other deliveries; Figure). Further, 34.5% of the women with a periviable birth had previous term deliveries only. Among these, the incidences of 1, 2, 3, and R4 consecutive previous term births were 53.0%, 25.2%, 11.7%, and 10.3%, respec-
tively. In aggregate, 78.9% of women with a periviable birth had no previous preterm birth to identify them as being at high risk for subsequent periviable birth. Causes of admission before periviable birth were labor (36.3%), PROM (34.1%), vaginal bleeding (10.4%), fetal death (6.2%), preeclampsia (4.0%), fetal abnormalities (2.4%), incompetent cervix (0.5%), fetal growth restriction (0.2%), and other complications (2.0%); 4% were twin/multifetal gestations (Table I.) We identified 7970 women who delivered both their first and second pregnancies at R20 weeks of gestation in our institution. Periviable birth in the first pregnancy was associated strongly with subsequent periviable birth and an increased risk of with preterm delivery at 20 to 30, 20 to 34, and 20 to 36 weeks (P ! .0001) for each (Table II). Periviable birth was associated most highly with periviable birth in the second pregnancy (relative risk, 8.6). This association decreased but remained significant when less severe preterm birth in the second pregnancy was considered (P ! .0001). Women with a periviable birth had a 35.6% risk of subsequent preterm birth and a 6.9% risk of recurrent periviable birth. Periviable birth in the first pregnancy was insensitive,
1178
Mercer, Milluzzi, and Collin
Table III Predictive value of preterm birth at 20 to 26, 20 to 30, 20 to 34, and 20 to 36 weeks of gestation in the first pregnancy for periviable birth at 20 to 26 weeks in the second pregnancy Gestation at delivery in first pregnancy Variable
20-26 Wk
20-30 Wk
20-34 Wk
30-36 Wk
Preterm birth in first pregnancy No previous preterm birth P value Relative risk of preterm birth in the second pregnancy Sensitivity of preterm birth for subsequent periviable birth
6.9 0.8 !.0001 8.6 8.8
5.7 0.73 !.0001 7.7 17.6
4.2 0.42 !.0001 6.6 30.9
2.5 0.25 !.0001 5.6 36.8
Table IV Time from admission to delivery and cervical dilation at admission for 919 women who were delivered at 20 to 26 weeks of gestation between 1994 and 2004 Indication for admission Preterm labor Premature rupture of the membranes Twins/multifetal gestations Vaginal bleeding Preeclampsia
Time to delivery (h)
Cervix (cm)
13.7 (21.5) 15.4 (35.5)
4.4 (2.8) 3.4 (2.8)
14.1 (24.4) 13.2 (17.6) 29.5 (31.6)
4.6 (3.3) 4.1 (3.3) 0.7 (0.9)
Data are given as mean (SD).
identifying only 8.8% of recurrent periviable births. Because of the poor predictive value and insensitivity of periviable birth, we evaluated associations between preterm birth at various gestational ages and subsequent periviable birth (Table III). A strong association was identified for each (P ! .0001), and this association was strongest with periviable birth in the first pregnancy. However, even preterm birth before 37 weeks of gestation was insensitive (sensitivity, 36.8%) for subsequent periviable birth. We found time from admission to delivery to be relatively brief (n = 919) and that many women had advanced cervical dilation, regardless of indication for admission.(Table IV). On average, women who were delivered at 20 to 26 weeks of gestation were in the hospital 13 to 15 hours before delivery. However, women with preeclampsia with early cervical dilatation (0.7 G 0.9 cm) remained undelivered for nearly 30 hours, on average.
Comment We have evaluated the frequency of and the causes for admission before periviable birth and have studied the associations between periviable birth and subsequent pregnancy outcomes. We have also evaluated the anticipated latency from admission to delivery for women who were delivered of a periviable infant because of various pregnancy complications. Periviable birth complicates
approximately 2% of the pregnancies at our tertiary-care institution, with approximately 1 in 3 of these complications occurring before the limit of potential viability. This likely over-estimates the general population risk, because women who are destined to deliver a periviable infant are referred preferentially to tertiary care institutions with facilities for critically ill infants. Not only is transfer reasonable, but also it is often feasible. Women who are delivered at 20 to 26 weeks of gestation remain pregnant 13 to 15 hours after admission, with a significant fraction of the women having extended latencies. The time that is available is often adequate to initiate other effective treatments (such as antenatal corticosteroid and GBS prophylaxis administration), neonatology consultation, and family counseling regarding anticipated outcomes and available interventions. The published literature varies considerably regarding the proximate causes of early preterm births. We studied primary indications for admission because that is more likely to be documented clearly in the patient record than subsequent complications that led to delivery. It is the condition that is evident on arrival and is the object of initial obstetric intervention. The indication for admission likely correlates closely to delivery indication when delivery occurs soon thereafter. Alternatively, some women will be admitted for one reason (eg, preterm labor or PROM) and could subsequently be delivered for other indications (eg, chorioamnionitis, membrane rupture, fetal distress or death, abruption, or labor), particularly if latency is prolonged. Currently, there are no data available regarding symptoms that are present before admission and that might be amenable to preventative treatment. Preterm birth is more common among twin/multifetal gestations, which are increasingly common as assisted reproductive technologies become more widely available. In the study of Bottoms et al,5 237 of 1045 infants were derived from twin/multifetal gestations. Although twin/multifetal gestations are represented disproportionately among periviable deliveries in our analysis (1/25 pregnancies; O1/12 periviable births), assisted reproductive technologies has not been available widely to our population over the past 30 years. We anticipate that these pregnancies will be even more prevalent
Mercer, Milluzzi, and Collin among periviable births in populations with more accessible assisted reproductive technologies. It is well-established that preterm birth is a risk factor for subsequent adverse outcomes.15-18 We found women with periviable births to be 3.4-fold more likely to have R2 previous abortions and more likely to have a previous preterm birth. Women with a periviable birth in the first pregnancy were 8.6-fold and 3.3-fold more likely to have subsequent periviable birth (6.9% vs 0.8%) and preterm birth (35.6% vs 10.9%), respectively. Alternatively, it is striking that 79% of women with a periviable birth have either no previous delivery or have a history of term deliveries only. Only 1 in 5 women with a periviable birth has a history of preterm birth, and some of these women will also have had term births subsequently. Previous pregnancy outcome is insensitive for subsequent periviable birth and carries a low positive predictive value. Because of this, previous pregnancy outcome is unlikely to be a useful predictor of periviable birth. This, in combination with the brief latency from admission to delivery, highlights the importance of the identification of markers that will be evident early in pregnancy or even before conception. Currently, available ancillary technologies may not be helpful as cervical shortening does not usually occur in the first or early second trimester, and fetal fibronectin is often present in the vagina !20 weeks of gestation. Because there is a stronger correlation between early preterm birth and subsequent preterm birth and periviable, it is possible that these women may be inherently different from those who are delivered at or near term and that these genetic, physiologic or biologic characteristics may be evident either before or early in pregnancy. This study is not without limitations. Databases possess fundamental weaknesses, which include a lack of prospective data collection, a reliance on data that are available in the medical record, and the need for judgments in data abstractions. We restricted our analyses to variables that are critical to clinical decisionmaking and/or can be anticipated to be documented routinely in the hospital record. To reduce potential biases regarding the prediction of subsequent adverse outcomes that are introduced by antecedent or intervening pregnancies and selective patient recall, we studied only those women who were delivered of their first and second gestation at our institution. It is reassuring that the associations that were seen between first and second pregnancies in this study were consistent with those that were identified in previous analyses.12-15 The optimal method to identify predictors of periviable birth would be to follow a women prospectively from before or early pregnancy. However, because of the rarity of periviable births, many thousands of pregnancies would need to be studied. In summary, periviable birth is most commonly preceded by labor, PROM, antepartum hemorrhage,
1179 and preeclampsia; and the time from admission to deliver is generally brief but frequently adequate for transfer or intervention to enhance perinatal outcomes. Alternatively, cervical dilation is often advanced on admission, and delivery is commonly inevitable. Previous pregnancy outcome is highly associated with periviable birth but is insensitive in the prediction of subsequent outcomes. It is plausible that some women with a periviable birth will have a chronic predisposition to recurrent preterm birth. Further efforts are needed to identify such chronic markers for early preterm birth that might be evident early in pregnancy or even before conception.
References 1. Hamilton BE, Martin JA, Sutton PD. Births: preliminary data for 2002: national vital statistics reports from the Centers for Disease Control And Prevention, National Center For Health Statistics. Natl Vital Stat 2003;51:1-20. 2. Demissie K, Rhoads GG, Ananth CV, Alexander GR, Kramer MS, Kogan MD, et al. Trends in preterm birth and neonatal mortality among blacks and whites in the United States from 1989 to 1997. Am J Epidemiol 2001;154:307-15. 3. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2002. Natl Vital Stat Rep 2003;52:1-114. 4. Stevenson DK, Wright LL, Lemons JA, Oh W, Korones SB, Papile LA, et al. Very low birth weight outcomes of the NICHD Neonatal Research Network, January 1993 through December 1994. Am J Obstet Gynecol 1998;179:1632-9. 5. Bottoms SF, Paul RH, Mercer BM, MacPherson CA, Roberts JM, Caritis SN, et al, for the NICHD Network of Maternal Fetal Medicine Research Units. Obstetrical determinants of neonatal survival: antenatal predictors of neonatal survival and morbidity in extremely low birth weight infants. Am J Obstet Gynecol 1999;180:665-9. 6. American College of Obstetricians and Gynecologists. Perinatal care at the threshold of viability. Washington (DC): The College; 2002. Clinical Management Guidelines no.: 38. 7. Wise PH, Wampler N, Barfield W. The importance of extreme prematurity and low birth weight to US neonatal mortality patterns: implications for prenatal care and women’s health. J Am Med Womens Assoc 1995;50:152-5. 8. Mercer BM. Premature rupture of the membranes. Obstet Gynecol 2003;101:178-93. 9. Fanaroff AA, Wright LL, Stevenson DK, Shankaran S, Donovan EF, Ehrenkranz RA, et al. Very-low-birth-weight outcomes of the NICHD Neonatal Research Network, May 1991 through December 1992. Am J Obstet Gynecol 1995;173:1423-31. 10. Tyson JE, Younes N, Verter J, Wright LL. Viability, morbidity, and resource use among newborns of 501- to 800-g birth weight: NICHD Neonatal Research Network. JAMA 1996;276:1645-51. 11. Louis JM, Ehrenberg HM, Collin MF, Mercer BM. Perinatal intervention and neonatal outcomes near the limit of viability. Am J Obstet Gynecol 2004;191:1398-402. 12. Mercer BM, Goldenberg RL, Das AF, Moawad AH, Iams JD, Meis PJ, et al and the NICHD MFMU Network. The preterm prediction study: a clinical risk assessment system. Am J Obstet Gynecol 1996;174:1885-95. 13. Adams MM, Elam Evans LD, Wilson HG, Gilberts DA. Rates and factors associated with recurrence of preterm delivery. JAMA 2000;283:1591-6.
1180 14. Mercer BM, Goldenberg RL, Moawad AH, Meis PJ, Iams JD, Das AF, et al. The preterm prediction study: effect of gestational age and cause of preterm birth on subsequent obstetric outcome. Am J Obstet Gynecol 1999;181:1216-21. 15. Mercer BM, Goldenberg RL, Meis PJ, Moawad AH, Shellhaas C, Das AF, et al. The Preterm Prediction Study: prediction of preterm premature rupture of membranes through clinical findings and ancillary testing: the NICHD Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol 2000;183:738-45. 16. Fliegner JR. Can anything be done about mid-trimester fetal wastage? Aust N Z J Obstet Gynaecol 1987;27:205-9. 17. Tucker JM, Goldenberg RL, Davis RO, Copper RL, Winkler CL, Hauth JC. Etiologies of preterm birth in an indigent
Mercer, Milluzzi, and Collin population: is prevention a logical expectation? Obstet Gynecol 1991;77:343-7. 18. Gray PH, Hurley TM, Rogers YM, O’Callaghan MJ, Tudehope DI, Burns YR, et al. Survival and neonatal and neurodevelopmental outcome of 24-29 week gestation infants according to primary cause of preterm delivery. Aust N Z J Obstet Gynaecol 1997;37:161-8. 19. Brenner WE, Edelman DA, Hendricks CH. A standard of fetal growth for the United States of America. Am J Obstet Gynecol 1976;126:555-64. 20. Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17: 857-72.
www.ajog.org
Periviable birth at 20 to 26 weeks of gestation: Proximate causes, previous obstetric history and recurrence risk Brian Mercer, MD,a Cynthia Milluzzi, BFA, BSN,a Marc Collin, MDb Departments of Reproductive Biologya and Pediatrics,b MetroHealth Medical Center at Case Western Reserve University, Cleveland, OH Received for publication March 1, 2005; revised May 2, 2005; accepted May 9, 2005
KEY WORDS Preterm birth Periviable birth Prematurity Prediction
Objective: Early preterm birth at 20 to 26 weeks of gestation (periviable birth) carries extreme risks of infant death and morbidities. Prevention of periviable birth could improve infant outcomes significantly. We sought to characterize the causes of periviable birth and to determine whether periviable birth can be predicted by previous pregnancy outcome. Study design: We evaluated 104,921 pregnancies (1974-2004) and assessed the frequency and causes of periviable birth. Women who were delivered of both their first and second pregnancies at O20 weeks of gestation at our institution were identified. Predictive values of the first pregnancy outcomes for second pregnancy outcomes were determined. Results: Periviable birth complicated 1981 deliveries (1.9%). Seventy-nine percent of the women with periviable births had no history of periviable births; 44% of the women had no previous deliveries, and 35% of the women had previous term deliveries only. Causes of periviable birth were labor (36%), premature rupture of membranes (34%), bleeding (10%), and preeclampsia (4%). Four percent of the gestations were multiple gestations. Among 7970 pregnancies at O20 weeks of gestation, periviable birth in the first pregnancy was associated with preterm birth and periviable birth in the second pregnancy (35.6%, 6.9%; relative risk, 3.3 and 8.6; P ! .0001). Periviable birth and preterm birth in the first pregnancy were insensitive for periviable birth in the second pregnancy (8.8%, 36.8%, respectively). Conclusion: Although periviable birth is associated with subsequent periviable birth and preterm birth, preterm birth and periviable birth are insensitive markers for recurrences in the next pregnancy. Early pregnancy or preconceptional markers for prediction of periviable birth are needed. Ó 2005 Mosby, Inc. All rights reserved.
Despite efforts directed towards prevention of preterm birth, the incidence of prematurity has increased by one-third from 9% to 12% over the past 2 decades and
Presented at the 25th Annual Meeting of the Society for Maternal Fetal Medicine, February 7-12, 2005, Reno, Nevada. Reprints not available from the authors. 0002-9378/$ - see front matter Ó 2005 Mosby, Inc. All rights reserved. doi:10.1016/j.ajog.2005.05.040
remains twice as common among African American births as white births.1-3 Most perinatal death results from complications of prematurity. Long-term sequelae that include neurologic handicaps, blindness, deafness, and chronic respiratory disease are linked directly to preterm birth, particularly before 30 weeks of gestation.4-6 Very low birth weight infants (!1500 g) account for 64% of all neonatal deaths.7
1176 Preterm birth that occurs near the limit of potential viability carries profound implications for the infant, family, and society. Survival is not anticipated with delivery before 23 weeks. Thereafter, survival increases progressively and steeply with each week that is gained between 23 and 30 weeks of gestation.8,9 With extreme prematurity, most survivors experience significant acute morbidities and are at risk for long-term sequelae.5,9,10 However, acute and subacute perinatal morbidities decrease with relatively small increases in gestational age at delivery from 23 weeks of gestation.7-11 Women suffering a preterm birth are at increased risk for recurrence in subsequent gestations.12 This risk increases with an increasing number of and decreasing gestational age at previous preterm births.12-15 Further, the risk of early preterm birth also increases with decreasing gestational age of the earliest previous preterm birth.14 However, studies regarding early preterm birth have disagreed as to the prevalence of preterm labor, premature rupture of the membranes, and other causes.5,16-18 The prevention of early preterm birth offers the greatest potential to reduce infant morbidity and mortality rates. Acute interventions to prevent preterm birth are not highly effective once symptoms occur. Because of this, we undertook this study to characterize the causes of admission before early preterm birth and to determine the ability of previous outcome to predict subsequent pregnancy early preterm birth.
Material and methods This retrospective cohort study was approved by the Institutional Review Board of MetroHealth Medical Center, Case Western Reserve University. We evaluated data that were collected between December 1974 and June 2004 by chart review at discharge and stored in our computer-based perinatal database. We evaluated all pregnancies that delivered a live- or stillborn infant at R20 weeks of gestation. Those infants with birth weights !150 g were considered implausible and were excluded.19 Our analysis was restricted to elements that had been collected continuously over the 30-year existence of this database and included gravidity, parity, admission indication, admission/delivery date and time, admission cervical dilation, delivery gestation (caregivers’ best obstetric estimate), and birth weight. Admission indications included labor, preterm rupture of membranes (PROM), incompetent cervix, vaginal bleeding, preeclampsia, fetal abnormalities, fetal death or growth restriction, and ‘‘other complications.’’ Twin/multifetal pregnancies were included once for each pregnancy and were considered separately. We defined periviable birth as a delivery that occurred between 20 weeks 0 days of gestation and 26
Mercer, Milluzzi, and Collin
Figure Previous obstetric outcome among women who were delivered at 20 weeks to 26 weeks 6 days of gestation in the current pregnancy.
weeks 6 days of gestation, with a birth weight between 150 and 1499 g.19 All pregnancies that met the inclusion criteria were evaluated to determine the frequency of periviable birth, and admission indications were determined for this subgroup. We then selected, from the aforementioned cohort, only those women who were delivered of both their first and second pregnancy at R20 weeks of gestation at our institution (G2P2s). Women with any miscarriages or therapeutic abortions and women with an intervening pregnancy between the first and second pregnancy were excluded, as were women with twin/ multifetal gestations in either pregnancy. We determined the likelihood of preterm and periviable delivery in the second pregnancy on the basis of the presence or absence of a similar history in the first pregnancy. We subsequently evaluated the time from admission to delivery for women with a periviable birth to determine the opportunity for intervention to prevent delivery or maximize fetal/neonatal outcomes. This information was calculated from the times and dates of admission and delivery (date, hour, minutes), which was documented in the database between January 1994 and June 2004. Descriptive statistical analyses were performed with Statview software (version 5.0.1; SAS Institute Inc, Cary, NC). Ninety-five percent confidence intervals for point estimates were determined by the method of Wilson.20 Fisher’s exact test was used to compare outcome frequencies. A probability value of !.05 was considered significant.
Results Of 104,921 pregnancies that were delivered between December 1974 and June 2004, periviable birth at 20
Mercer, Milluzzi, and Collin Table I
1177
Point estimates for various causes of early preterm birth in 5 studies
Inclusion criteria
Fliegner et al12 (1987) 22-28 wk
Tucker et al13 (1991) 20-29 wk
Gray et al14 (1997) 24-29 wk
Bottoms et al5 (1999) %1000 g
Mercer et al (2005) 20-26 wk
Sample size (n) Premature labor (n) Premature rupture of the membranes (n) Antepartum hemorrhage (n) Fetal death (n) Hypertension/preeclampsia (n) Multiple pregnancy (n) Congenital malformations (n) Incompetent cervix (n) Fetal growth restriction (n) Chorioamnionitis (n) Fetal distress (n) Other or unexplained (n)
d* 18 12 13 d* 7 14 11 13 d* d* d* 12
354 23 33 4 13 2 d* 2 d* 4 2 2 17
189 14 34 20 d* 14 d* d* d* d* 8 d* 10
713 42 22 4 d* 12 d* d*
1981 36.3 (34.2-28.4)y 34.1 (32.0-36.2)y 10.4 (9.1-11.8)y 6.2 (5.2-7.4)y 4.0 (3.2-4.9)y 4.0 (3.2-4.9)y 2.4 (1.8-3.2)y 0.5 (0.3-0.9)y 0.2 (0.08-0.5)y d* d* 2.0 (1.5-2.7)y
2 d* 9 9
* Not evaluated. y Data given as percent (95% CI).
Table II Predictive value of periviable birth at 20 to 26 weeks of gestation in the first pregnancy for preterm birth at 20 to 26, 20 to 30, 20 to 34, and 20 to 36 weeks of gestation in the second pregnancy Gestation at delivery in second pregnancy Variable
20-26 Wk
20-30 Wk
20-34 Wk
30-36 Wk
Periviable birth in first pregnancy (%) No previous periviable birth (%) Relative risk of periviable birth in the second pregnancy P value Sensitivity of periviable birth for subsequent preterm birth (%)
6.9 0.8 8.6 !.0001 8.8
11.5 1.8 6.4 !.0001 6.8
19.5 4.9 4.0 !.0001 4.5
35.6 10.9 3.3 !.0001 3.5
weeks to 26 weeks 6 days of gestation complicated 1981 deliveries (1.9%). One third of these deliveries (33.1%) occurred before the limit of viability (20 weeks-22 weeks 6 days of gestation). The incidences of preterm birth at 20 to 30, 20 to 34, and 20 to 36 weeks of gestation were 4.0%, 9.3%, and 16.2%, respectively. Women with periviable births were much more likely to have had R2 previous spontaneous or induced abortions than those without a periviable birth (23.8% vs 7.1%; P ! .0001). Although more likely to have had a previous preterm birth than women without such a history (21.1% vs 15.6%; P ! .0001), only 1in 5 women with a periviable birth had a history of R1 previous preterm births. Among these, 52.6% of the women had R1 previous term deliveries in addition to the previous preterm birth. Forty-four percent of the women had no previous deliveries (25.5% were in their first pregnancy, and 18.9% had R1 previous miscarriages and/or therapeutic abortions but no other deliveries; Figure). Further, 34.5% of the women with a periviable birth had previous term deliveries only. Among these, the incidences of 1, 2, 3, and R4 consecutive previous term births were 53.0%, 25.2%, 11.7%, and 10.3%, respec-
tively. In aggregate, 78.9% of women with a periviable birth had no previous preterm birth to identify them as being at high risk for subsequent periviable birth. Causes of admission before periviable birth were labor (36.3%), PROM (34.1%), vaginal bleeding (10.4%), fetal death (6.2%), preeclampsia (4.0%), fetal abnormalities (2.4%), incompetent cervix (0.5%), fetal growth restriction (0.2%), and other complications (2.0%); 4% were twin/multifetal gestations (Table I.) We identified 7970 women who delivered both their first and second pregnancies at R20 weeks of gestation in our institution. Periviable birth in the first pregnancy was associated strongly with subsequent periviable birth and an increased risk of with preterm delivery at 20 to 30, 20 to 34, and 20 to 36 weeks (P ! .0001) for each (Table II). Periviable birth was associated most highly with periviable birth in the second pregnancy (relative risk, 8.6). This association decreased but remained significant when less severe preterm birth in the second pregnancy was considered (P ! .0001). Women with a periviable birth had a 35.6% risk of subsequent preterm birth and a 6.9% risk of recurrent periviable birth. Periviable birth in the first pregnancy was insensitive,
1178
Mercer, Milluzzi, and Collin
Table III Predictive value of preterm birth at 20 to 26, 20 to 30, 20 to 34, and 20 to 36 weeks of gestation in the first pregnancy for periviable birth at 20 to 26 weeks in the second pregnancy Gestation at delivery in first pregnancy Variable
20-26 Wk
20-30 Wk
20-34 Wk
30-36 Wk
Preterm birth in first pregnancy No previous preterm birth P value Relative risk of preterm birth in the second pregnancy Sensitivity of preterm birth for subsequent periviable birth
6.9 0.8 !.0001 8.6 8.8
5.7 0.73 !.0001 7.7 17.6
4.2 0.42 !.0001 6.6 30.9
2.5 0.25 !.0001 5.6 36.8
Table IV Time from admission to delivery and cervical dilation at admission for 919 women who were delivered at 20 to 26 weeks of gestation between 1994 and 2004 Indication for admission Preterm labor Premature rupture of the membranes Twins/multifetal gestations Vaginal bleeding Preeclampsia
Time to delivery (h)
Cervix (cm)
13.7 (21.5) 15.4 (35.5)
4.4 (2.8) 3.4 (2.8)
14.1 (24.4) 13.2 (17.6) 29.5 (31.6)
4.6 (3.3) 4.1 (3.3) 0.7 (0.9)
Data are given as mean (SD).
identifying only 8.8% of recurrent periviable births. Because of the poor predictive value and insensitivity of periviable birth, we evaluated associations between preterm birth at various gestational ages and subsequent periviable birth (Table III). A strong association was identified for each (P ! .0001), and this association was strongest with periviable birth in the first pregnancy. However, even preterm birth before 37 weeks of gestation was insensitive (sensitivity, 36.8%) for subsequent periviable birth. We found time from admission to delivery to be relatively brief (n = 919) and that many women had advanced cervical dilation, regardless of indication for admission.(Table IV). On average, women who were delivered at 20 to 26 weeks of gestation were in the hospital 13 to 15 hours before delivery. However, women with preeclampsia with early cervical dilatation (0.7 G 0.9 cm) remained undelivered for nearly 30 hours, on average.
Comment We have evaluated the frequency of and the causes for admission before periviable birth and have studied the associations between periviable birth and subsequent pregnancy outcomes. We have also evaluated the anticipated latency from admission to delivery for women who were delivered of a periviable infant because of various pregnancy complications. Periviable birth complicates
approximately 2% of the pregnancies at our tertiary-care institution, with approximately 1 in 3 of these complications occurring before the limit of potential viability. This likely over-estimates the general population risk, because women who are destined to deliver a periviable infant are referred preferentially to tertiary care institutions with facilities for critically ill infants. Not only is transfer reasonable, but also it is often feasible. Women who are delivered at 20 to 26 weeks of gestation remain pregnant 13 to 15 hours after admission, with a significant fraction of the women having extended latencies. The time that is available is often adequate to initiate other effective treatments (such as antenatal corticosteroid and GBS prophylaxis administration), neonatology consultation, and family counseling regarding anticipated outcomes and available interventions. The published literature varies considerably regarding the proximate causes of early preterm births. We studied primary indications for admission because that is more likely to be documented clearly in the patient record than subsequent complications that led to delivery. It is the condition that is evident on arrival and is the object of initial obstetric intervention. The indication for admission likely correlates closely to delivery indication when delivery occurs soon thereafter. Alternatively, some women will be admitted for one reason (eg, preterm labor or PROM) and could subsequently be delivered for other indications (eg, chorioamnionitis, membrane rupture, fetal distress or death, abruption, or labor), particularly if latency is prolonged. Currently, there are no data available regarding symptoms that are present before admission and that might be amenable to preventative treatment. Preterm birth is more common among twin/multifetal gestations, which are increasingly common as assisted reproductive technologies become more widely available. In the study of Bottoms et al,5 237 of 1045 infants were derived from twin/multifetal gestations. Although twin/multifetal gestations are represented disproportionately among periviable deliveries in our analysis (1/25 pregnancies; O1/12 periviable births), assisted reproductive technologies has not been available widely to our population over the past 30 years. We anticipate that these pregnancies will be even more prevalent
Mercer, Milluzzi, and Collin among periviable births in populations with more accessible assisted reproductive technologies. It is well-established that preterm birth is a risk factor for subsequent adverse outcomes.15-18 We found women with periviable births to be 3.4-fold more likely to have R2 previous abortions and more likely to have a previous preterm birth. Women with a periviable birth in the first pregnancy were 8.6-fold and 3.3-fold more likely to have subsequent periviable birth (6.9% vs 0.8%) and preterm birth (35.6% vs 10.9%), respectively. Alternatively, it is striking that 79% of women with a periviable birth have either no previous delivery or have a history of term deliveries only. Only 1 in 5 women with a periviable birth has a history of preterm birth, and some of these women will also have had term births subsequently. Previous pregnancy outcome is insensitive for subsequent periviable birth and carries a low positive predictive value. Because of this, previous pregnancy outcome is unlikely to be a useful predictor of periviable birth. This, in combination with the brief latency from admission to delivery, highlights the importance of the identification of markers that will be evident early in pregnancy or even before conception. Currently, available ancillary technologies may not be helpful as cervical shortening does not usually occur in the first or early second trimester, and fetal fibronectin is often present in the vagina !20 weeks of gestation. Because there is a stronger correlation between early preterm birth and subsequent preterm birth and periviable, it is possible that these women may be inherently different from those who are delivered at or near term and that these genetic, physiologic or biologic characteristics may be evident either before or early in pregnancy. This study is not without limitations. Databases possess fundamental weaknesses, which include a lack of prospective data collection, a reliance on data that are available in the medical record, and the need for judgments in data abstractions. We restricted our analyses to variables that are critical to clinical decisionmaking and/or can be anticipated to be documented routinely in the hospital record. To reduce potential biases regarding the prediction of subsequent adverse outcomes that are introduced by antecedent or intervening pregnancies and selective patient recall, we studied only those women who were delivered of their first and second gestation at our institution. It is reassuring that the associations that were seen between first and second pregnancies in this study were consistent with those that were identified in previous analyses.12-15 The optimal method to identify predictors of periviable birth would be to follow a women prospectively from before or early pregnancy. However, because of the rarity of periviable births, many thousands of pregnancies would need to be studied. In summary, periviable birth is most commonly preceded by labor, PROM, antepartum hemorrhage,
1179 and preeclampsia; and the time from admission to deliver is generally brief but frequently adequate for transfer or intervention to enhance perinatal outcomes. Alternatively, cervical dilation is often advanced on admission, and delivery is commonly inevitable. Previous pregnancy outcome is highly associated with periviable birth but is insensitive in the prediction of subsequent outcomes. It is plausible that some women with a periviable birth will have a chronic predisposition to recurrent preterm birth. Further efforts are needed to identify such chronic markers for early preterm birth that might be evident early in pregnancy or even before conception.
References 1. Hamilton BE, Martin JA, Sutton PD. Births: preliminary data for 2002: national vital statistics reports from the Centers for Disease Control And Prevention, National Center For Health Statistics. Natl Vital Stat 2003;51:1-20. 2. Demissie K, Rhoads GG, Ananth CV, Alexander GR, Kramer MS, Kogan MD, et al. Trends in preterm birth and neonatal mortality among blacks and whites in the United States from 1989 to 1997. Am J Epidemiol 2001;154:307-15. 3. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2002. Natl Vital Stat Rep 2003;52:1-114. 4. Stevenson DK, Wright LL, Lemons JA, Oh W, Korones SB, Papile LA, et al. Very low birth weight outcomes of the NICHD Neonatal Research Network, January 1993 through December 1994. Am J Obstet Gynecol 1998;179:1632-9. 5. Bottoms SF, Paul RH, Mercer BM, MacPherson CA, Roberts JM, Caritis SN, et al, for the NICHD Network of Maternal Fetal Medicine Research Units. Obstetrical determinants of neonatal survival: antenatal predictors of neonatal survival and morbidity in extremely low birth weight infants. Am J Obstet Gynecol 1999;180:665-9. 6. American College of Obstetricians and Gynecologists. Perinatal care at the threshold of viability. Washington (DC): The College; 2002. Clinical Management Guidelines no.: 38. 7. Wise PH, Wampler N, Barfield W. The importance of extreme prematurity and low birth weight to US neonatal mortality patterns: implications for prenatal care and women’s health. J Am Med Womens Assoc 1995;50:152-5. 8. Mercer BM. Premature rupture of the membranes. Obstet Gynecol 2003;101:178-93. 9. Fanaroff AA, Wright LL, Stevenson DK, Shankaran S, Donovan EF, Ehrenkranz RA, et al. Very-low-birth-weight outcomes of the NICHD Neonatal Research Network, May 1991 through December 1992. Am J Obstet Gynecol 1995;173:1423-31. 10. Tyson JE, Younes N, Verter J, Wright LL. Viability, morbidity, and resource use among newborns of 501- to 800-g birth weight: NICHD Neonatal Research Network. JAMA 1996;276:1645-51. 11. Louis JM, Ehrenberg HM, Collin MF, Mercer BM. Perinatal intervention and neonatal outcomes near the limit of viability. Am J Obstet Gynecol 2004;191:1398-402. 12. Mercer BM, Goldenberg RL, Das AF, Moawad AH, Iams JD, Meis PJ, et al and the NICHD MFMU Network. The preterm prediction study: a clinical risk assessment system. Am J Obstet Gynecol 1996;174:1885-95. 13. Adams MM, Elam Evans LD, Wilson HG, Gilberts DA. Rates and factors associated with recurrence of preterm delivery. JAMA 2000;283:1591-6.
1180 14. Mercer BM, Goldenberg RL, Moawad AH, Meis PJ, Iams JD, Das AF, et al. The preterm prediction study: effect of gestational age and cause of preterm birth on subsequent obstetric outcome. Am J Obstet Gynecol 1999;181:1216-21. 15. Mercer BM, Goldenberg RL, Meis PJ, Moawad AH, Shellhaas C, Das AF, et al. The Preterm Prediction Study: prediction of preterm premature rupture of membranes through clinical findings and ancillary testing: the NICHD Maternal-Fetal Medicine Units Network. Am J Obstet Gynecol 2000;183:738-45. 16. Fliegner JR. Can anything be done about mid-trimester fetal wastage? Aust N Z J Obstet Gynaecol 1987;27:205-9. 17. Tucker JM, Goldenberg RL, Davis RO, Copper RL, Winkler CL, Hauth JC. Etiologies of preterm birth in an indigent
Mercer, Milluzzi, and Collin population: is prevention a logical expectation? Obstet Gynecol 1991;77:343-7. 18. Gray PH, Hurley TM, Rogers YM, O’Callaghan MJ, Tudehope DI, Burns YR, et al. Survival and neonatal and neurodevelopmental outcome of 24-29 week gestation infants according to primary cause of preterm delivery. Aust N Z J Obstet Gynaecol 1997;37:161-8. 19. Brenner WE, Edelman DA, Hendricks CH. A standard of fetal growth for the United States of America. Am J Obstet Gynecol 1976;126:555-64. 20. Newcombe RG. Two-sided confidence intervals for the single proportion: comparison of seven methods. Stat Med 1998;17: 857-72.