Group B streptococci (GBS) is a bacterium that colonizes the vagina or rectum in roughly 25% of healthy, adult women and in pregnancy can be transmitted to the neonate.1,2 Intrapartum antibiotic prophylaxis for GBS-colonized women has resulted in a significant reduction in early-onset neonatal GBS infection.3 Cesarean delivery does not prevent mother-to-child transmission of GBS because GBS can cross intact amniotic membranes, but when cesarean delivery is performed before the onset of labor on a woman with intact membranes, the risk for early-onset GBS disease among full-term neonates is extremely low.4 The 2010 Centers for Disease Control and Prevention GBS guidelines recommend that women planning cesarean delivery should undergo routine screening for GBS at 35–37 weeks of gestation because onset of labor or rupture of membranes can occur before planned cesarean delivery. Under those circumstances, GBS-colonized women should receive intrapartum antibiotic prophylaxis.4
The cesarean delivery rate in 2014 in the United States was 32.2%.5 Repeat cesarean deliveries comprise approximately 40% of all cesarean deliveries6 and most women do not labor before the scheduled date of delivery.7,8 Additionally, of those presenting in labor, many will undergo cesarean delivery before membrane rupture. The objective of this study was therefore to estimate the cost-effectiveness of the current standard of care, universal GBS screening in women with a singleton pregnancy intending to have a repeat cesarean delivery compared with no routine GBS screening.
MATERIALS AND METHODS
We developed a decision model to evaluate the costs and effects of two strategies for GBS screening in women with a prior cesarean delivery and a current singleton pregnancy planning to undergo a repeat cesarean delivery (universal GBS screening compared with no GBS screening) in a theoretical cohort of 400,000 pregnant women in the United States (the approximate total number of repeat cesarean deliveries performed per year in the United States) at 35 weeks of gestation. In the universal screening group, a rectovaginal swab was performed to screen for GBS at 35 weeks of gestation. In the no screening group, no woman had a rectovaginal swab performed for GBS. The analysis was performed from a health care perspective to estimate the total expenditures related to GBS screening and treatment.
In developing the model, we assumed that all women in the universal screening group underwent screening at 35 weeks of gestation. Women who were screened and found to be GBS-positive received intrapartum antibiotic prophylaxis if they presented in labor before their scheduled cesarean delivery regardless of the eventual mode of delivery. Women whose GBS status was unknown, either because they were not screened or because the culture was not yet resulted (a proportion of those tested at 35 weeks of gestation who then presented in labor during that week, because a GBS culture takes 2–3 days to result), received intrapartum antibiotic prophylaxis based on risk-based criteria (ie, antibiotics if less than 37 weeks of gestation). We assumed that, if a women presented in labor before her scheduled cesarean delivery, her neonate was at risk of early-onset GBS disease regardless of status of the membranes (intact compared with ruptured), duration of rupture of membranes, time from presentation to cesarean delivery, or final mode of delivery. We assumed some women underwent a prelabor cesarean delivery before 39 weeks of gestation for medical or obstetric indications. In the model, a delivery during each week beyond 35 weeks of gestation but before 39 weeks of gestation could be a prelabor repeat cesarean delivery for medical or obstetric complications, a repeat cesarean delivery in the setting of labor, or a vaginal birth after cesarean delivery.
Neonates were assumed to not be at risk of early-onset GBS disease if a prelabor cesarean delivery was performed (Ramus R, McIntire D, Wendell G. Antibiotic chemoprophylaxis for group B strep is not necessary with elective cesarean section at term [Abstract]. Am J Obstet Gynecol 1999;180[suppl]:85.),9,10 although this was varied in sensitivity analyses. Late-onset GBS disease was not included in the model because it was assumed to be equal in both groups because intrapartum antibiotic prophylaxis has not been shown to be effective in its reduction.3 The proportion of women who presented in labor and received intrapartum antibiotic prophylaxis was the same whether they had a vaginal or cesarean delivery. This antibiotic prophylaxis was assumed to have a similar effectiveness regardless of mode of delivery3,10; however, given limited data, this was also varied in sensitivity analysis. All women who underwent cesarean delivery received prophylactic antibiotics for the cesarean delivery; however, given the paucity of data on the subject and the short duration of neonatal antibiotic exposure in this setting, we assumed that pre–cesarean delivery antibiotic prophylaxis did not reduce the risk of early-onset GBS disease. Early-onset GBS disease resulted in increased cost and a prolonged neonatal hospital stay.
To obtain base case probability point estimates and confidence intervals, we conducted an English language search of PubMed to identify relevant publications. The final search terms included: pregnancy, vaginal birth after cesarean, cesarean section, Streptococcus agalactiae, and cost. The search was not limited by publication date or country of origin. All identified documents were examined and those that were relevant retrieved. Reference lists of retrieved documents were manually reviewed to identify additional publications. Point estimates were determined from published randomized controlled trials, prospective cohorts, and national vital statistic data when possible. Retrospective cohorts or review studies were used when no other sources of information were available. If there was not one study that was methodologically superior, we calculated base case point estimates as the unweighted mean or median of the available data based on their distributions (Table 1).
Based on published literature, we assumed a GBS prevalence of 25% (range 10–50%), that 26.6% of women would labor before their scheduled cesarean delivery (range 13.5–35%), and that overall, only 3.3% of women who planned a repeat cesarean delivery would instead deliver vaginally (range 0–50%).
We derived utilities from published literature. Utilities are a means of evaluating the relative quality of life as compared with health. We determined five neonatal health states that would be relevant for our analysis: normal health (utility=1), severe disability (utility=0.48), moderate disability (utility=0.69), mild disability (utility=0.89), and neonatal death (utility=0).11–13 The utilities were derived using the Health Utilities Index and EuroQol system11–13 and author judgment.11 Severe disability is defined as serious medical conditions that significantly limit working capacity and include quadriplegia, blindness, deafness, hydrocephalus, uncontrolled epilepsy, or severe learning disability. Moderate disability includes medical conditions such as diplegia, hemiplegia, or moderate learning disability. Mild disability includes medical conditions such as myopia, language delay, mild hearing loss, or hyperactivity.11,14 We assigned an average life expectancy of 79 years for healthy nonates and for those with mild disability, 68 years for neonates with moderate disability, and 26 years for neonates with severe disability.11–13 We did not model the prolonged neonatal hospital stay as a neonatal disutility. To calculate quality-adjusted life-years (QALYs), we assumed a discounting rate of 3% (range 0–5%) (Table 2). To determine the exact QALY value, the utility value associated with a given state of health was multiplied by the years lived in that state. Discounting assumes that current health is worth more than future health meaning that the utility of each subsequent year is decreased by 3% in the QALY calculation.
We derived cost estimates in a similar fashion to the probability estimates, but additionally queried local and national hospital and insurance data (Table 2). We adjusted all costs to reflect 2015 U.S. dollars. The costs accounted for in the model included the cost of testing, maternal follow-up and treatment, neonatal care, and the lifetime cost of a disability. Offspring costs were broken down into neonatal care costs and long-term disability costs. Long-term care costs included only direct medical expenses; thus, productivity losses were not included.
The primary outcome was the cost per neonatal QALY gained with a willingness to pay of $100,000 per neonatal QALY gained, a standard used threshold in cost-effectiveness analysis.15–17 In addition to the base case analysis, we performed one-way sensitivity analyses and Monte Carlo simulation. Monte Carlo simulation is a computational algorithm that relies on repeated random sampling of all variables across their confidence intervals based on their distributions. The distributions used in the Monte Carlo simulation were triangular or normal for the probabilities and γ for the costs. A triangular distribution has fixed minimum, most likely, and maximum values and has a probability distribution function that appears triangular and is not necessarily symmetric. The normal distribution is symmetric and does not have fixed minimum and maximum values, but rather utilizes the standard deviation of the parameter to approximate probabilities. The γ distribution is similar to the normal distribution, but is right-skewed, which allows for higher costs to be sampled in the Monte Carlo simulation. Given the plausible variation in all of the probability, cost, and utility estimates included in the model, no variable was excluded from the Monte Carlo analysis. One hundred thousand simulations were run to estimate the percentage of time that universal GBS screening would be cost-effective as compared with no screening. We performed all analyses using TreeAge Pro 2016 Suite. The study did not involve human participants and was exempt from institutional review board approval.
In the base case, we assumed there to be a GBS prevalence of 25%, that 26.6% of women would labor between 35 weeks of gestation and their scheduled cesarean delivery, and that overall, 3.3% women would deliver vaginally. Universal GBS screening in women with a singleton pregnancy intending to have a repeat cesarean delivery was not cost-effective compared with no screening, costing $114,445 per neonatal QALY gained. The number of women needed to be screened to prevent one case of early-onset GBS was 28,442, costing $426,666. The number of women needed to be screened to prevent one neonatal death was 1,154,027, costing $17,312,185. Finally, the number of women needed to be screened to prevent any GBS-related disability was 111,874, costing $1,678,274. Assuming base case estimates, of every 400,000 women screened (the approximate total number of repeat cesarean deliveries per year in the United States), 14 cases of early-onset GBS, no neonatal deaths, and three GBS-related disabilities would be prevented.
In sensitivity analyses, the model was most sensitive to the prevalence of maternal GBS colonization, the percentage of women who labored before their scheduled cesarean delivery and the percentage of women who ultimately delivered vaginally (Fig. 1). Universal screening became cost-effective with relatively small changes to these variables: if greater than 28% of women were GBS-positive, greater than 29% of women labored before their scheduled repeat cesarean delivery or greater than 10% ultimately delivered vaginally. Examples of high-risk populations such as obese women or women who were colonized with GBS in a prior pregnancy (therefore, women at high risk for GBS colonization) or women with a prior preterm delivery (therefore at risk of delivery before a scheduled cesarean delivery) are presented in Table 3 with the respective costs per neonatal QALY. Because costs and QALYs are especially important when building models that evaluate the lifespan of a neonate, additional one-way sensitivity analyses were performed to investigate these variables (Fig. 2). This diagram demonstrates that the quality of life attributed to each health state affected the cost-effectiveness of the intervention significantly more than the lifetime cost of each health state.
Using Monte Carlo Simulation, universal GBS screening was cost-saving in 50% of simulations and was cost-effective at a threshold of $100,000 per neonatal QALY in an additional 22%. Overall, universal GBS screening was cost-effective or cost-saving in 72% of all possible scenarios.
Universal GBS screening in women with a singleton pregnancy intending to have a repeat cesarean delivery may not be cost-effective in all populations. In our base case, it would cost more than $100,000 in direct costs per neonatal QALY gained. However, GBS screening may be cost-effective in many other scenarios; for example, in populations with a high GBS prevalence, in women at high risk of laboring before their scheduled cesarean delivery, and in women who may ultimately opt for a vaginal delivery. When evaluating all possible scenarios using Monte Carlo simulation, universal screening was cost-effective more than 70% of the time.
Our base case results suggest that universal screening is not cost-effective, whereas our Monte Carlo results indicate that universal screening is cost-effective in a majority of plausible situations. This discordance may be the result of inherent bias in our model. When data were unknown, we chose to allow the plausible situations to favor GBS screening. For example, only a proportion of women who present in labor before their scheduled cesarean delivery also have ruptured membranes. It is much less likely to transmit GBS with intact membranes, therefore making the overall probability of early-onset GBS disease lower in this proportion of neonates. Second, we did not take into account the fact that all women undergoing cesarean delivery receive preoperative antibiotics, which may treat some GBS disease. Although, given the short duration of antibiotic exposure in this scenario, this effect may be negligible. Third, not all women who are GBS-colonized or GBS unknown with risk factors who present in labor before a planned cesarean delivery actually receive antibiotics for GBS regardless of having been tested. This is the result of not only time constraints, but also health care provider oversight. Treating all of these women in the decision tree made screening more likely cost-effective.
Strengths of this study are the ability to evaluate large ranges of costs and probabilities. We were additionally able to account for both short- and long-term costs and QALYs associated with early-onset GBS disease. The major limitation of all cost analyses is that the evaluation is only as good as the available data and generally based on population averages. For example, there is only one study that specifically evaluated the number of women who labor before a scheduled cesarean delivery8; and the actual percentage of women who present with ruptured membranes and receive antibiotic prophylaxis for GBS before cesarean delivery is unknown. Additionally, the length of time a woman waits for her cesarean delivery after presenting with ruptured membranes is unknown and likely affects the rate of GBS transmission and also the chance of receiving antibiotic prophylaxis. Furthermore, the overall risk of early-onset GBS disease in women who present with ruptured membranes or in labor and then undergo cesarean delivery is unknown regardless of the timing of cesarean delivery. We additionally did not consider multiple gestations in this analysis. There are very limited data evaluating timing of labor in women planning a repeat cesarean delivery for twins (or higher order multiples) as well as risk of GBS disease in twins before repeat cesarean delivery. Also, the current literature advocates for earlier delivery of twins in general,18 which will mostly prevent labor before delivery.
Another limitation is that outcomes such as a neonatal death led to a neonatal QALY of zero. This skews the interpretation of the cost-effectiveness ratio because a severe disability, in which a QALY is assigned and utilized in the analysis, may be perceived as a worse outcome than death. We also did not account for maternal QALYs because there are no published data on this topic.
We did not account for indirect costs such as loss of work, and it is conceivable that we underestimated costs, especially the cost related to morbidity from GBS disease. Additionally, we assumed that each factor evaluated by one-way sensitivity analysis is independent. However, if a woman colonized with GBS is more likely to deliver vaginally or labor before a scheduled cesarean delivery, the independence assumption is not valid and the cost-effectiveness is altered for that group of women. Finally, no model can adequately address the potential downstream effect of changing a guideline. If we stop universal GBS screening for women planning a repeat cesarean delivery, it is possible that health care providers would accidentally forget to screen other populations such as women desiring a trial of labor after cesarean delivery. This complexity was not accounted for in the model.
Based on the results of this analysis, we are concerned that universal screening may not be cost-effective in women who are having scheduled repeat cesarean deliveries. However, given the many plausible scenarios in which universal screening is cost-effective, we would recommend continued universal screening for GBS at this time. Further studies should evaluate the effectiveness of antibiotic therapy for GBS in the setting of labor before cesarean delivery, local adherence to the recommendation to give antibiotic prophylaxis to GBS-colonized women in labor planning cesarean delivery, and the effectiveness of routine preoperative antibiotics to prevent GBS transmission. Only with more detailed data can we determine whether universal GBS screening is cost-effective for the entire U.S. population.
1. Regan JA, Klebanoff MA, Nugent RP. The epidemiology of group B streptococcal colonization in pregnancy. Vaginal Infections and Prematurity Study Group. Obstet Gynecol 1991;77:604–10.
2. Haberland CA, Benitz WE, Sanders GD, Pietzsch JB, Yamada S, Nguyen L, et al. Perinatal screening for group B streptococci: cost-benefit analysis of rapid polymerase chain reaction. Pediatrics 2002;110:471–80.
3. Ohlsson A, Shah VS. Intrapartum antibiotics for known maternal Group B streptococcal colonization. The Cochrane Database of Systematic Reviews 2014, Issue 6. Art. No.: CD007467. doi: 10.1002/14651858.CD007467.pub4.
4. Verani JR, McGee L, Schrag SJ; Division of Bacterial Diseases, National Center for Immunization and Respiratory Diseases, Centers for Disease Control and Prevention (CDC). Prevention of perinatal group B streptococcal disease—revised guidelines from CDC, 2010. MMWR Recomm Rep 2010;59:1–36.
5. Hamilton BE, Martin JA, Osterman MJK, Curtin SC, Matthews TJ. Births: final data for 2014. Natl Vital Stat Rep 2015;64:1–64.
6. Osterman MJ, Martin JA. Primary cesarean delivery rates, by state: results from the revised birth certificate, 2006-2012. Natl Vital Stat Rep 2014;63:1–11.
7. Landon MB, Hauth JC, Leveno KJ, Spong CY, Leindecker S, Varner MW, et al. Maternal and perinatal outcomes associated with a trial of labor after prior cesarean delivery. N Engl J Med 2004;351:2581–9.
8. Roberts CL, Nicholl MC, Algert CS, Ford JB, Morris JM, Chen JS. Rate of spontaneous onset of labour before planned repeat caesarean section at term. BMC Pregnancy Childbirth 2014;14:125.
9. Berardi A, Rossi K, Pizzi C, Baronciani D, Venturelli C, Ferrari F, et al. Absence of neonatal streptococcal colonization after planned cesarean section. Acta Obstet Gynecol Scand 2006;85:1012–3.
10. Håkansson S, Axemo P, Bremme K, Bryngelsson AL, Wallin MC, Ekström CM, et al. Group B streptococcal carriage in Sweden: a national study on risk factors for mother and infant colonisation. Acta Obstet Gynecol Scand 2008;87:50–8.
11. Tengs TO, Wallace A. One thousand health-related quality-of-life estimates. Med Care 2000;38:583–637.
12. Oostenbrink R, A Moll HA, Essink-Bot ML. The EQ-5D and the Health Utilities Index for permanent sequelae after meningitis: a head-to-head comparison. J Clin Epidemiol 2002;55:791–9.
13. Colbourn T, Asseburg C, Bojke L, Philips Z, Claxton K, Ades AE, et al. Prenatal screening and treatment strategies to prevent group B streptococcal and other bacterial infections in early infancy: cost-effectiveness and expected value of information analyses. Health Technol Assess 2007;11:1–226, iii.
14. Moster D, Lie RT, Markestad T. Long-term medical and social consequences of preterm birth. N Engl J Med 2008;359:262–73.
15. Hirth RA, Chernew ME, Miller E, Fendrick AM, Weissert WG. Willingness to pay for a quality-adjusted life year: in search of a standard. Med Decis Making 2000;20:332–42.
16. Kaplan RM, Bush JW. Health-related quality of life measurement for evaluation research and policy analysis. Health Psychol 1982;1:61–80.
17. Werner EF. Creating decision trees to assess cost-effectiveness in clinical research. J Biomet Biostat 2012. Available at: http://www.omicsonline.org/2155-6180/2155-6180-S7-004.digital/2155-6180-S7-004.html
. Retrieved September 21, 2016.
18. Dodd JM, Deussen AR, Grivell RM, Crowther CA. Elective birth at 37 weeks' gestation for women with an uncomplicated twin pregnancy. The Cochrane Database of Systematic Reviews 2014, Issue 2. Art. No.: CD003582. doi: 10.1002/14651858.CD003582.pub2.
19. Turrentine MA, Colicchia LC, Hirsch E, Cheng PJ, Tam T, Ramsey PS, et al. Efficiency of screening for the recurrence of antenatal group B streptococcus colonization in a subsequent pregnancy: a systematic review and meta-analysis with independent patient data. Am J Perinatol 2016;33:510–7.
20. Kleweis SM, Cahill AG, Odibo AO, Tuuli MG. Maternal obesity and rectovaginal group B streptococcus colonization at term. Infect Dis Obstet Gynecol 2015;2015:586767.
21. Stapleton RD, Kahn JM, Evans LE, Critchlow CW, Gardella CM. Risk factors for group B streptococcal genitourinary tract colonization in pregnant women. Obstet Gynecol 2005;106:1246–52.
22. Koumans EH, Rosen J, van Dyke MK, Zell E, Phares CR, Taylor A, et al. Prevention of mother-to-child transmission of infections during pregnancy: implementation of recommended interventions, United States, 2003–2004. Am J Obstet Gynecol 2012;206:158.e1–11.
23. Page-Ramsey SM, Johnstone SK, Kim D, Ramsey PS. Prevalence of group B streptococcus colonization in subsequent pregnancies of group B streptococcus-colonized versus noncolonized women. Am J Perinatol 2013;30:383–8.
24. Martin JA, Hamilton BE, Ventura SJ, Osterman MJK, Kirmeyer S, Mathews TJ, et al. Births: final data for 2009. Natl Vital Stat Rep 2011;60:1–70.
25. Martin JA, Hamilton BE, Osterman MJ, Curtin SC, Matthews TJ. Births: final data for 2013. Natl Vital Stat Rep 2015;64:1–65.
26. Slattery MM, Morrison JJ. Preterm delivery. Lancet 2002;360:1489–97.
27. Davies HD, Miller MA, Faro S, Gregson D, Kehl SC, Jordan JA. Multicenter study of a rapid molecular-based assay for the diagnosis of group B streptococcus colonization in pregnant women. Clin Infect Dis 2004;39:1129–35.
28. Gavino M, Wang E. A comparison of a new rapid real-time polymerase chain reaction system to traditional culture in determining group B streptococcus colonization. Am J Obstet Gynecol 2007;197:388.e1–4.
29. Money D, Dobson S, Cole L, Karacabeyli E, Blondel-Hill E, Milner R, et al. An evaluation of a rapid real time polymerase chain reaction assay for detection of group B streptococcus as part of a neonatal group B streptococcus prevention strategy. J Obstet Gynaecol Can 2008;30:770–5.
30. Young BC, Dodge LE, Gupta M, Rhee JS, Hacker MR. Evaluation of a rapid, real-time intrapartum group B streptococcus assay. Am J Obstet Gynecol 2011;205:372.e1–6.
31. Crowther CA, Dodd JM, Hiller JE, Haslam RR, Robinson JS; Birth After Caesarean Study Group. Planned vaginal birth or elective repeat caesarean: patient preference restricted cohort with nested randomised trial. PLoS Med 2012;9:e1001192.
32. Schrag SJ, Verani JR. Intrapartum antibiotic prophylaxis for the prevention of perinatal group B streptococcal disease: experience in the United States and implications for a potential group B streptococcal vaccine. Vaccine 2013;31(suppl 4):D20–6.
33. Paccione KA, Wiesenfeld HC. Guideline adherence for intrapartum group B streptococci prophylaxis in penicillin-allergic patients. Infect Dis Obstet Gynecol 2013;2013:917304.
34. Critchfield AS, Lievense SP, Raker CA, Matteson KA. Group B streptococcus prophylaxis in patients who report a penicillin allergy: a follow-up study. Am J Obstet Gynecol 2011;204:150.e1–8.
35. Fairlie T, Zell ER, Schrag S. Effectiveness of intrapartum antibiotic prophylaxis for prevention of early-onset group B streptococcal disease. Obstet Gynecol 2013;121:570–7.
36. Briody VA, Albright CM, Has P, Hughes BL. Use of cefazolin for group B streptococci prophylaxis in women reporting a penicillin allergy without anaphylaxis. Obstet Gynecol 2016;127:577–83.
37. Apter AJ, Kinman JL, Bilker WB, Herlim M, Margolis DJ, Lautenbach E, et al. Represcription of penicillin after allergic-like events. J Allergy Clin Immunol 2004;113:764–70.
38. Chang C, Mahmood MM, Teuber SS, Gershwin ME. Overview of penicillin allergy. Clin Rev Allergy Immunol 2012;43:84–97.
39. Johannes CB, Ziyadeh N, Seeger JD, Tucker E, Reiter C, Faich G. Incidence of allergic reactions associated with antibacterial use in a large, managed care organisation. Drug Saf 2007;30:705–13.
40. Tice AD, Rehm SJ, Dalovisio JR, Bradley JS, Martinelli LP, Graham DR, et al. Practice guidelines for outpatient parenteral antimicrobial therapy. IDSA guidelines. Clin Infect Dis 2004;38:1651–72.
41. Oster G, Edelsberg J, Hennegan K, Lewin C, Narasimhan V, Slobod K, et al. Prevention of group B streptococcal disease in the first 3 months of life: would routine maternal immunization during pregnancy be cost-effective? Vaccine 2014;32:4778–85.
42. Centers for Disease Control and Prevention (CDC). Trends in perinatal group B streptococcal disease—United States, 2000-2006. MMWR Morb Mortal Wkly Rep 2009;58:109–12.
43. Turrentine MA, Ramirez MM, Mastrobattista JM. Cost-effectiveness of universal prophylaxis in pregnancy with prior group B streptococci colonization. Infect Dis Obstet Gynecol 2009;2009:934698.
44. Phibbs CS, Schmitt SK. Estimates of the cost and length of stay changes that can be attributed to one-week increases in gestational age for premature infants. Early Hum Dev 2006;82:85–95.
45. El Helali N, Giovangrandi Y, Guyot K, Chevet K, Gutmann L, Durand-Zaleski I. Cost and effectiveness of intrapartum group B streptococcus polymerase chain reaction screening for term deliveries. Obstet Gynecol 2012;119:822–9.
46. Waitzman NJ, Romano PS, Scheffler RM. Estimates of the economic costs of birth defects. Inquiry 1994;31:188–205.
47. Centers for Disease Control and Prevention (CDC). Economic costs associated with mental retardation, cerebral palsy, hearing loss, and vision impairment—United States, 2003. MMWR Morb Mortal Wkly Rep 2004;53:57–9.
48. Bloom SL, Yost NP, McIntire DD, Leveno KJ. Recurrence of preterm birth in singleton and twin pregnancies. Obstet Gynecol 2001;98:379–85.
49. Spong CY. Prediction and prevention of recurrent spontaneous preterm birth. Obstet Gynecol 2007;110:405–15.