- Email: [email protected]
Placental edge to internal os distance in the late third trimester and mode of delivery in placenta praevia
BJOG: an International Journal of Obstetrics and Gynaecology September 2003, Vol. 110, pp. 860– 864
Placental edge to internal os distance in the late third trimester and mode of delivery in placenta praevia Amarnath Bhide, Federico Prefumo*, Jessica Moore, Brian Hollis, Basky Thilaganathan Objectives To correlate transvaginal ultrasound findings with mode of delivery in cases of placenta praevia. Design Cohort study. Setting A London Teaching Hospital. Methods Retrospective review of all cases of placenta praevia diagnosed by transvaginal ultrasound between February 1997 and March 2002. Main outcome measures Likelihood of vaginal delivery and major obstetric haemorrhage. Results A total of 121 pregnancies were studied with a mean scan-to-delivery interval of 10.5 days. In the 64 women who laboured, the likelihood of vaginal delivery rose significantly as the placental edge to internal os distance increased. Caesarean section rate was 90% when the placental edge – internal os distance was 0.1 to 2.0 cm, falling to 37% when this measurement was over 2.0 cm ( P < 0.00045). Conclusion Trial of vaginal delivery is appropriate in cases with a placental to internal os distance >2 cm. The term ‘praevia’ should be restricted to cases where the placental edge is 2 cm from the internal os, as the likelihood of operative delivery and significant postpartum haemorrhage is high. Cases where the placenta is more than 2 cm from the internal os have a greater than 60% chance of vaginal delivery and should be defined as ‘low lying’ in order to reduce the clinician’s bias towards operative delivery. INTRODUCTION Antepartum haemorrhage complicates 2– 5% of pregnancies of which approximately one-third are due to placenta praevia1. Traditionally, placenta praevia is defined as the implantation of the placenta in the lower uterine segment. The severity of placenta praevia is determined by the relationship of the placental edge to the lower uterine segment. Originally, this was assessed clinically, and there were established management protocols for the various clinical types. Ultrasound has now become the gold standard for the evaluation of cases with suspected placenta praevia2. Despite the widespread use of ultrasound to diagnose placenta praevia, there are no well designed, large scale studies available to base either an ultrasound classification or subsequent clinical management. The two most relevant studies involved few women, variable criteria for diagnosis, ultrasound examinations performed early in the third trimester and retrospectively defined cohorts3,4. These studies have been of limited clinical value in establishing ultrasound criteria for diagnosis of placenta praevia and
Fetal Medicine Unit, Academic Department of Obstetrics and Gynaecology, St George’s Hospital Medical School, London, UK * Correspondence: Dr F. Prefumo, Fetal Medicine Unit, St George’s Hospital Medical School, Cranmer Terrace, London SW17 0RE, UK. D RCOG 2003 BJOG: an International Journal of Obstetrics and Gynaecology PII: S 1 4 7 0 - 0 3 2 8 ( 0 3 ) 0 2 9 9 1 - 4
determining the appropriate mode of delivery. The aim of the present study is to determine the mode of delivery in women with placenta praevia diagnosed on third trimester ultrasound.
METHODS This is a retrospective analysis of all cases of placenta praevia booked for routine antenatal care and delivery in a single unit between February 1997 and March 2002. All cases were diagnosed by transvaginal ultrasound, which was performed for a number of reasons including: identification of a low lying at the time of the routine 21 –22 week anomaly scan, antepartum haemorrhage or abnormal presentation in the third trimester of pregnancy. The patients were managed according to the departmental protocol. All cases where the placenta was low lying on the anomaly scan were booked to have placental localisation from 32 weeks onwards. A suspected diagnosis of placenta praevia on abdominal ultrasound was further assessed by transvaginal ultrasound. The hospital policy is to admit all cases of placenta praevia presenting with antepartum haemorrhage. Women are allowed to go home if they do not have fresh bleeding for a period of over one week after the admission. However, re-admission with a repeat bleeding episode was an indication for admission till delivery. Elective caesarean section for placenta praevia was scheduled for 38 –39 weeks. An emergency caesarean section was performed if women with major praevia went www.bjog-elsevier.com
PLACENTA PRAEVIA AND MODE OF DELIVERY
861
Table 1. Demographic and delivery characteristics of the 121 women with placenta praevia. Cases are grouped according to the distance between placental edge and internal cervical os. All the values are expressed as median (range). Group 1 (Overlap or reaching os) No. of cases Age (years) Gestational age at last scan (weeks þ days) Gestational age at delivery (weeks þ days) Interval between scan and delivery (days) Birthweight (g)
32 36 þ 1 37 þ 5 11 2965
42 (19 to 42) (33 þ 1 to 38 þ 5) (33 þ 6 to 39 þ 0) (0 to 34) (2085 to 4200)
into preterm labour or developed significant vaginal bleeding. The mode of delivery in cases of minor placenta praevia cases was decided on clinical grounds. Missed diagnosis of placenta praevia is a recognised clinical risk trigger and was reported and investigated routinely. The ultrasound examinations were performed after any active vaginal bleeding had ceased and directly after voiding. Transvaginal ultrasound was performed using ATL HDI 5000, ATL HDI 3000 (Letchworth, UK) or Acuson XP128 (Mountain View, California, USA) ultrasound machines equipped with a transvaginal probe (5.0 – 7.5 MHz). The probe, covered with sterile cover, was carefully inserted in the lower part of the vagina till the cervix came into view. A sagittal scan of the whole length of the cervix with the lower part of the uterus was obtained. The probe was then rotated through 90j in either direction, keeping the cervical canal constantly in view. The image was assessed for the presence of the lower placental edge. A major praevia was diagnosed if placental tissue extended to or covered the internal cervical os. Where the placental edge was visible but did not cross the internal os, the smallest intervening distance was measured. We only included in the analysis cases where the lower edge of the placenta was 3.5 cm or less from the internal cervical os. The cases of placenta praevia were identified by a computer search on the obstetric ultrasound database. Only the most recent scan prior to delivery was included into the analysis. Obstetric outcome was determined by reviewing the antenatal and delivery notes for all patients. Delivery was defined as elective caesarean section, if surgery was performed at a planned date on an elective operative list. All other caesarean sections were classified as emergency ones. The latter were divided into those performed for presumed bleeding from the placenta praevia or for other
Group 2 (0.1 – 2.0 cm)
34 36 þ 4 38 þ 2 11.5 3130
40 (20 to 45) (32 þ 6 to 39 þ 0) (35 þ 0 to 40 þ 4) (1 to 32) (1900 to 3800)
Group 3 (2.1 – 3.5 cm)
33 36 þ 4 38 þ 6 15 3090
39 (21 to 42) (33 þ 1 to 40 þ 1) (33 þ 2 to 41 þ 2) (0 to 47) (1900 to 3800)
obstetric reasons. Postpartum haemorrhage was defined as an estimated blood loss > 500 mL in women delivered vaginally. Because the average blood loss at a caesarean section is about 1000 mL, a loss >1000 mL was considered a postpartum haemorrhage in women undergoing a caesarean section5. Statistical analysis was performed using the software SPSS for Windows, release 10.0.5 (SPSS, Chicago, Illinois, USA). Mann – Whitney U test, m2 test and Fisher’s exact test were used where appropriate.
RESULTS The entry criteria were fulfilled in 125 pregnancies. Follow up was not available in four women who delivered in other units; hence, 121 cases were available for analysis. The cases were divided into three group for analysis. Group 1 consisted of women were the placenta edge reached or overlapped the internal os. Group 2 included those women where the placental edge was from 0.1 to 2.0 cm from the internal os, while Group 3 included women where the placental edge was more than 2.0 cm from the internal os. The demographic characteristics of the three groups were similar (Table 1). There were no reported cases of missed diagnosis of placenta praevia through the risk management process during the study period. However, it is possible that some cases of minor placenta praevia that were not detected on ultrasound were allowed to labour and deliver, while remaining unreported. Table 2 shows the prevalence of antepartum and postpartum haemorrhage in the three groups. The trend for a higher rate of antepartum haemorrhage with increasing degree of placenta praevia was statistically significant
Table 2. Prevalence of antepartum haemorrhage and postpartum haemorrhage. The higher rates of antepartum haemorrhage with increasing degree of placenta were statistically significant (m2 ¼ 7.05, P ¼ 0.0294). Placental edge to internal os distance
Overlap or reaching os (Group 1) 0.1 – 2.0 cm (Group 2) 2.1 – 3.5 cm (Group 3) Total
No. of pregnancies
Delivery <34 weeks (n)
Delivery <37 weeks (n)
42 40 39 121
1 0 1 2
14 9 3 26
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
No. with antepartum haemorrhage (%) 24 19 11 54
(57.1) (47.5) (28.2) (44.6)
No. with postpartum haemorrhage (%) 5 2 3 10
(11.9) (5.0) (7.7) (8.3)
862
A. BHIDE ET AL.
Table 3. Likelihood of undergoing a caesarean section for any reason with an ultrasound diagnosis of placenta praevia (Fisher’s exact test: Group 1 vs 2, P ¼ 0.4704; Group 2 vs 3, P < 0.0001; Group 1 vs 3, P < 0.00005). The likelihood of vaginal delivery in labouring women based on the distance between placental edge and the internal cervical os is also shown. The difference in the proportion of women delivering vaginally between Group 2 and Group 3 is statistically significant (Fisher’s exact test, P < 0.0005). Placental edge – internal os distance
No.
Elective caesarean section
No. of women presenting in labour
Emergency caesarean section for bleeding in labour
Emergency caesarean section for other reasons
Vaginal delivery (% of labour)
Group 1 (overlap or reaching os) Group 2 (0.1 – 2.0 cm) Group 3 (2.1 – 3.5 cm)
42 40 39
25 20 12
17 20 27
10 6 2
7 12 8
0 (0) 2 (10) 17 (63)
(m2 ¼ 7.05, P < 0.005). Table 3 illustrates the rates of caesarean section and vaginal delivery in the three groups. The overall caesarean section rate is also shown graphically in groups of 1 cm intervals in Fig. 1.
DISCUSSION This is the largest study of women with placenta praevia diagnosed using transvaginal ultrasound in the late third trimester of pregnancy. The only two previous studies to examine ultrasound findings in late pregnancy included fewer women and had a larger time interval between ultrasound and delivery3,4. Oppenheimer et al.3 performed a retrospective analysis of 52 cases of placenta praevia where ultrasound examinations were performed, on average, five weeks prior to delivery. Dawson et al.4 used translabial ultrasound in 40 women with suspected placenta praevia with an average interval of six weeks between ultrasound and delivery. In the latter study, apart from the use of translabial rather than transvaginal ultrasound, only 31 women had a placenta that was less than 3.0 cm from the internal os at 32 weeks of gestation.
Fig. 1. Prevalence of caesarean section in labouring women according to the distance between placental edge and internal cervical os. Means and their 95% confidence intervals are displayed. The dotted line represents the background caesarean section rate (23.8%).
The time interval between ultrasound and delivery has a significant influence on the classification and clinical outcome of placenta praevia. Several studies have demonstrated that the distance between the placental edge and the internal cervical os changes with advancing gestation6 – 13. Indeed, in one study, the mean rate of placental migration was estimated at 0.54 cm per week in the third trimester of pregnancy13. Thus, in the current study with an average of two weeks from ultrasound to delivery, the placenta is unlikely to have migrated more than 1.0 cm. In contrast, the placenta is likely to have migrated, on average, 2.7 and 3.2 cm in the two previous studies where the ultrasound to delivery interval was five and six weeks, respectively. The upper limit of the placental edge to internal os distance for the use of the term ‘praevia’ is unclear and undefined. Oppenheimer et al.3 included cases where this distance was up to 5.8 cm. Dawson et al.4 arbitrarily chose an upper limit of 3.0 cm. In the present study, an upper limit of 3.5 cm was used, as none of the cases of antepartum haemorrhage with a placental edge to internal os distance of over 3.5 cm were suspected to be due to placenta praevia. We acknowledge that this definition is arbitrary and based on clinical interpretation of ultrasound findings. However, analysis of the current data set allows the establishment of guidelines for delivery. Given the paucity of published evidence, the Royal College of Obstetricians and Gynaecologists14 recommend that trial of vaginal delivery is appropriate if the placental edge – internal os distance is over 2.0 cm. This guideline formed the basis for our post hoc definition of placenta praevia groups. The Royal College of Obstetricians and Gynaecologists’ recommendation is based on the findings of Oppenheimer et al.3, who performed a retrospective analysis of placenta praevia, where the data were grouped according to mode of delivery rather than antenatal ultrasound findings. They reported a mean placental edge to internal os distance measured five weeks prior to delivery of 1.1 cm (range 0 to 2.0 cm) in cases requiring a caesarean section (n ¼ 7), as opposed to 3.1 cm (range 1.8– 5.8 cm) in those delivering vaginally (n ¼ 14). The findings of the latter study are of limited clinical value because the grouping based on mode of delivery does not provide useful information for clinicians who need to make decisions based D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
PLACENTA PRAEVIA AND MODE OF DELIVERY
on antenatal ultrasound findings. The probability of vaginal delivery at a given placental edge – internal os distance cannot be calculated from the latter study. Indeed, the authors themselves concluded that, given the small numbers, the clinical significance of their findings was to be interpreted with caution and further studies were needed to establish guidelines. In the present study, the cohorts for comparison were identified on the basis of antenatal ultrasound findings. The findings of this study are that the likelihood for successful vaginal delivery increases with the distance of the placental edge from the internal os. This is clearly demonstrated in Table 3 and is illustrated in Fig. 1. Even though the emergency caesarean section rate was 90% in Group 2 and 37% in Group 3, these figures are likely to be influenced by clinical bias. Knowledge of the placental location, which was available to the clinicians managing the case, would probably have played an important part in the decision to perform an operative delivery. Hence, vaginal delivery rates in this study are likely to be conservative estimates. Clinicians may have been happier to conduct a trial of labour if they had sufficient scientific evidence to support an expectant management strategy. When evaluating the likelihood of vaginal delivery, it is clear from the data that placenta praevia reaching or overlapping the internal cervical os requires elective caesarean delivery. It is equally evident that when the placental edge is greater than 2 cm from the internal os, the likelihood of a successful trial of vaginal delivery is good. In Group 2, where the placental edge to internal os distance was 0.1 to 2 cm, some women achieved a successful vaginal delivery. However, all women with a placental edge to internal os distance of 1 cm or less required a caesarean delivery. Figure 1 also demonstrates a trend for decreasing caesarean section rates with increasing placental edge to internal os distances. Given the retrospective nature of the study and the likely clinical bias, management of cases with a placenta to internal os distance of 1.1 to 2 cm is unclear. A prospective study in women with placental edge – internal os distance of 1.1 to 2 cm may be needed to address this issue. It may be appropriate to allow these women a trial of labour with suitable precautions to manage emergency operative delivery and intrapartum/postpartum haemorrhage. The data of this study demonstrate that the likelihood for antepartum haemorrhage falls significantly from 57% in Group 1 to 28% in Group 3. It is important to note that a significant proportion (43%) of women with a placenta encroaching or overlapping the internal os did not have bleeding in the antenatal period. These data are supported with the two-in-three risk of antepartum haemorrhage reported in older series, where the diagnosis of placenta praevia was made on clinical grounds15. All the cases of postpartum haemorrhage in the present series occurred in women who had undergone a caesarean section. Five out of the 10 cases of postpartum haemorrhage took place in the group with the placenta overlapping D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
863
the internal os. One of these five cases was a placenta accreta, with an estimated blood loss of 14 L, necessitating a hysterectomy. The postpartum haemorrhage rate in Groups 2 and 3 was similar to the background postpartum haemorrhage rate within our unit.
CONCLUSION The data of this study support the trial of vaginal delivery in cases with a placental to internal os distance >2 cm and an elective caesarean section when this distance is 1 cm or less. We propose that the term ‘praevia’ should be restricted to cases where the placental edge is 2 cm or less from the internal os, as the likelihood of operative delivery and significant postpartum haemorrhage remains high. Cases where the placenta is more than 2 cm from the internal os have a high chance of vaginal delivery and should be defined as ‘low lying’ rather than praevia, in order to reduce the bias towards operative delivery in these women.
Acknowledgements Dr Prefumo was supported by a Marie Curie Fellowship of the European Community programme Quality of Life under contract number QLGA-CT-2000-52145.
References 1. Conje JC, Taylor DJ. Bleeding in late pregnancy. In: James DK, Steer PJ, Weiner CP, Gonik B, editors. High Risk Pregnancy: Management Options, 2nd edition. Philadelphia: Saunders, 1999:111. 2. Oyelese Y. Placenta praevia and vasa praevia: time to leave the dark ages. Ultrasound Obstet Gynecol 2001;18:96 – 99. 3. Oppenheimer LW, Farine D, Knox Ritchie JW, Lewinsky RM, Telford J, Fairbanks LA. What is a low-lying placenta? Am J Obstet Gynecol 1991;165:1036 – 1038. 4. Dawson W, Dumas M, Romano W, Gagnon R, Gratton R, Mowbray D. Translabial ultrasonography and placenta previa: does measurement of the os – placenta distance predict outcome? J Ultrasound Med 1996; 15:441 – 446. 5. Obstetric hemorrhage. In: Cunningham FG, MacDonald PC, Gant NF, et al., editors. Williams Obstetrics, 20th edition. Stamford, Connecticut: Appleton and Lange, 1997:745 – 782. 6. Ruparelia BA, Chapman MG. Early low-lying placentae — ultrasonic assessment, progress and outcome. Eur J Obstet Gynecol Reprod Biol 1986;20:209 – 213. 7. Townsend RR, Laing FC, Nyberg DA, Brook JR, Wing VW. Technical factors responsible for ‘placental migration’: sonographic assessment. Radiology 1986;160:105 – 108. 8. Mustafa SA, Brizot ML, Carvalho MH, Watanabe L, Kanhale S, Zugaib M. Transvaginal ultrasonography in predicting placenta previa at delivery: a longitudinal study. Ultrasound Obstet Gynecol 2002;20: 356 – 359. 9. Lauria MR, Smith RS, Treadwelll MC, et al. The use of secondtrimester transvaginal sonography to predict placenta previa. Ultrasound Obstet Gynecol 1996;8:337 – 340. 10. Taipale P, Hiilesmaa V, Ylostalo P. Transvaginal ultrasonography at
864
A. BHIDE ET AL.
18 – 23 weeks in predicting placenta previa at delivery. Ultrasound Obstet Gynecol 1998;12:422 – 425. 11. Gallagher P, Fagan CJ, Bedi DG, Winsett MZ, Reyes RN. Potential placenta previa: definition, frequency, and significance. Am J Roentgenol 1987;149:1013 – 1015. 12. Dashe JS, McIntire DD, Ramus RM, Santos-Ramos R, Twickler DM. Persistence of placenta previa according to gestational age at ultrasound detection. Obstet Gynecol 2002;99:692 – 697. 13. Oppenheimer L, Simpson P, Holmes N, Dabrowski A. Diagnosis of
low-lying placenta: can migration in the third trimester predict outcome? Ultrasound Obstet Gynecol 2001;18:100 – 102. 14. Placenta Praevia: Diagnosis and Management. Clinical Green-top guideline no. 27. London: The Royal College of Obstetricians and Gynaecologists, 2001 (January). 15. Hill DJ, Beischer NA. Placenta praevia without anterpartum haemorrhage. Aust N Z J Obstet Gynaecol 1980;20:21 – 23. Accepted 28 May 2003
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
Placental edge to internal os distance in the late third trimester and mode of delivery in placenta praevia Amarnath Bhide, Federico Prefumo*, Jessica Moore, Brian Hollis, Basky Thilaganathan Objectives To correlate transvaginal ultrasound findings with mode of delivery in cases of placenta praevia. Design Cohort study. Setting A London Teaching Hospital. Methods Retrospective review of all cases of placenta praevia diagnosed by transvaginal ultrasound between February 1997 and March 2002. Main outcome measures Likelihood of vaginal delivery and major obstetric haemorrhage. Results A total of 121 pregnancies were studied with a mean scan-to-delivery interval of 10.5 days. In the 64 women who laboured, the likelihood of vaginal delivery rose significantly as the placental edge to internal os distance increased. Caesarean section rate was 90% when the placental edge – internal os distance was 0.1 to 2.0 cm, falling to 37% when this measurement was over 2.0 cm ( P < 0.00045). Conclusion Trial of vaginal delivery is appropriate in cases with a placental to internal os distance >2 cm. The term ‘praevia’ should be restricted to cases where the placental edge is 2 cm from the internal os, as the likelihood of operative delivery and significant postpartum haemorrhage is high. Cases where the placenta is more than 2 cm from the internal os have a greater than 60% chance of vaginal delivery and should be defined as ‘low lying’ in order to reduce the clinician’s bias towards operative delivery. INTRODUCTION Antepartum haemorrhage complicates 2– 5% of pregnancies of which approximately one-third are due to placenta praevia1. Traditionally, placenta praevia is defined as the implantation of the placenta in the lower uterine segment. The severity of placenta praevia is determined by the relationship of the placental edge to the lower uterine segment. Originally, this was assessed clinically, and there were established management protocols for the various clinical types. Ultrasound has now become the gold standard for the evaluation of cases with suspected placenta praevia2. Despite the widespread use of ultrasound to diagnose placenta praevia, there are no well designed, large scale studies available to base either an ultrasound classification or subsequent clinical management. The two most relevant studies involved few women, variable criteria for diagnosis, ultrasound examinations performed early in the third trimester and retrospectively defined cohorts3,4. These studies have been of limited clinical value in establishing ultrasound criteria for diagnosis of placenta praevia and
Fetal Medicine Unit, Academic Department of Obstetrics and Gynaecology, St George’s Hospital Medical School, London, UK * Correspondence: Dr F. Prefumo, Fetal Medicine Unit, St George’s Hospital Medical School, Cranmer Terrace, London SW17 0RE, UK. D RCOG 2003 BJOG: an International Journal of Obstetrics and Gynaecology PII: S 1 4 7 0 - 0 3 2 8 ( 0 3 ) 0 2 9 9 1 - 4
determining the appropriate mode of delivery. The aim of the present study is to determine the mode of delivery in women with placenta praevia diagnosed on third trimester ultrasound.
METHODS This is a retrospective analysis of all cases of placenta praevia booked for routine antenatal care and delivery in a single unit between February 1997 and March 2002. All cases were diagnosed by transvaginal ultrasound, which was performed for a number of reasons including: identification of a low lying at the time of the routine 21 –22 week anomaly scan, antepartum haemorrhage or abnormal presentation in the third trimester of pregnancy. The patients were managed according to the departmental protocol. All cases where the placenta was low lying on the anomaly scan were booked to have placental localisation from 32 weeks onwards. A suspected diagnosis of placenta praevia on abdominal ultrasound was further assessed by transvaginal ultrasound. The hospital policy is to admit all cases of placenta praevia presenting with antepartum haemorrhage. Women are allowed to go home if they do not have fresh bleeding for a period of over one week after the admission. However, re-admission with a repeat bleeding episode was an indication for admission till delivery. Elective caesarean section for placenta praevia was scheduled for 38 –39 weeks. An emergency caesarean section was performed if women with major praevia went www.bjog-elsevier.com
PLACENTA PRAEVIA AND MODE OF DELIVERY
861
Table 1. Demographic and delivery characteristics of the 121 women with placenta praevia. Cases are grouped according to the distance between placental edge and internal cervical os. All the values are expressed as median (range). Group 1 (Overlap or reaching os) No. of cases Age (years) Gestational age at last scan (weeks þ days) Gestational age at delivery (weeks þ days) Interval between scan and delivery (days) Birthweight (g)
32 36 þ 1 37 þ 5 11 2965
42 (19 to 42) (33 þ 1 to 38 þ 5) (33 þ 6 to 39 þ 0) (0 to 34) (2085 to 4200)
into preterm labour or developed significant vaginal bleeding. The mode of delivery in cases of minor placenta praevia cases was decided on clinical grounds. Missed diagnosis of placenta praevia is a recognised clinical risk trigger and was reported and investigated routinely. The ultrasound examinations were performed after any active vaginal bleeding had ceased and directly after voiding. Transvaginal ultrasound was performed using ATL HDI 5000, ATL HDI 3000 (Letchworth, UK) or Acuson XP128 (Mountain View, California, USA) ultrasound machines equipped with a transvaginal probe (5.0 – 7.5 MHz). The probe, covered with sterile cover, was carefully inserted in the lower part of the vagina till the cervix came into view. A sagittal scan of the whole length of the cervix with the lower part of the uterus was obtained. The probe was then rotated through 90j in either direction, keeping the cervical canal constantly in view. The image was assessed for the presence of the lower placental edge. A major praevia was diagnosed if placental tissue extended to or covered the internal cervical os. Where the placental edge was visible but did not cross the internal os, the smallest intervening distance was measured. We only included in the analysis cases where the lower edge of the placenta was 3.5 cm or less from the internal cervical os. The cases of placenta praevia were identified by a computer search on the obstetric ultrasound database. Only the most recent scan prior to delivery was included into the analysis. Obstetric outcome was determined by reviewing the antenatal and delivery notes for all patients. Delivery was defined as elective caesarean section, if surgery was performed at a planned date on an elective operative list. All other caesarean sections were classified as emergency ones. The latter were divided into those performed for presumed bleeding from the placenta praevia or for other
Group 2 (0.1 – 2.0 cm)
34 36 þ 4 38 þ 2 11.5 3130
40 (20 to 45) (32 þ 6 to 39 þ 0) (35 þ 0 to 40 þ 4) (1 to 32) (1900 to 3800)
Group 3 (2.1 – 3.5 cm)
33 36 þ 4 38 þ 6 15 3090
39 (21 to 42) (33 þ 1 to 40 þ 1) (33 þ 2 to 41 þ 2) (0 to 47) (1900 to 3800)
obstetric reasons. Postpartum haemorrhage was defined as an estimated blood loss > 500 mL in women delivered vaginally. Because the average blood loss at a caesarean section is about 1000 mL, a loss >1000 mL was considered a postpartum haemorrhage in women undergoing a caesarean section5. Statistical analysis was performed using the software SPSS for Windows, release 10.0.5 (SPSS, Chicago, Illinois, USA). Mann – Whitney U test, m2 test and Fisher’s exact test were used where appropriate.
RESULTS The entry criteria were fulfilled in 125 pregnancies. Follow up was not available in four women who delivered in other units; hence, 121 cases were available for analysis. The cases were divided into three group for analysis. Group 1 consisted of women were the placenta edge reached or overlapped the internal os. Group 2 included those women where the placental edge was from 0.1 to 2.0 cm from the internal os, while Group 3 included women where the placental edge was more than 2.0 cm from the internal os. The demographic characteristics of the three groups were similar (Table 1). There were no reported cases of missed diagnosis of placenta praevia through the risk management process during the study period. However, it is possible that some cases of minor placenta praevia that were not detected on ultrasound were allowed to labour and deliver, while remaining unreported. Table 2 shows the prevalence of antepartum and postpartum haemorrhage in the three groups. The trend for a higher rate of antepartum haemorrhage with increasing degree of placenta praevia was statistically significant
Table 2. Prevalence of antepartum haemorrhage and postpartum haemorrhage. The higher rates of antepartum haemorrhage with increasing degree of placenta were statistically significant (m2 ¼ 7.05, P ¼ 0.0294). Placental edge to internal os distance
Overlap or reaching os (Group 1) 0.1 – 2.0 cm (Group 2) 2.1 – 3.5 cm (Group 3) Total
No. of pregnancies
Delivery <34 weeks (n)
Delivery <37 weeks (n)
42 40 39 121
1 0 1 2
14 9 3 26
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
No. with antepartum haemorrhage (%) 24 19 11 54
(57.1) (47.5) (28.2) (44.6)
No. with postpartum haemorrhage (%) 5 2 3 10
(11.9) (5.0) (7.7) (8.3)
862
A. BHIDE ET AL.
Table 3. Likelihood of undergoing a caesarean section for any reason with an ultrasound diagnosis of placenta praevia (Fisher’s exact test: Group 1 vs 2, P ¼ 0.4704; Group 2 vs 3, P < 0.0001; Group 1 vs 3, P < 0.00005). The likelihood of vaginal delivery in labouring women based on the distance between placental edge and the internal cervical os is also shown. The difference in the proportion of women delivering vaginally between Group 2 and Group 3 is statistically significant (Fisher’s exact test, P < 0.0005). Placental edge – internal os distance
No.
Elective caesarean section
No. of women presenting in labour
Emergency caesarean section for bleeding in labour
Emergency caesarean section for other reasons
Vaginal delivery (% of labour)
Group 1 (overlap or reaching os) Group 2 (0.1 – 2.0 cm) Group 3 (2.1 – 3.5 cm)
42 40 39
25 20 12
17 20 27
10 6 2
7 12 8
0 (0) 2 (10) 17 (63)
(m2 ¼ 7.05, P < 0.005). Table 3 illustrates the rates of caesarean section and vaginal delivery in the three groups. The overall caesarean section rate is also shown graphically in groups of 1 cm intervals in Fig. 1.
DISCUSSION This is the largest study of women with placenta praevia diagnosed using transvaginal ultrasound in the late third trimester of pregnancy. The only two previous studies to examine ultrasound findings in late pregnancy included fewer women and had a larger time interval between ultrasound and delivery3,4. Oppenheimer et al.3 performed a retrospective analysis of 52 cases of placenta praevia where ultrasound examinations were performed, on average, five weeks prior to delivery. Dawson et al.4 used translabial ultrasound in 40 women with suspected placenta praevia with an average interval of six weeks between ultrasound and delivery. In the latter study, apart from the use of translabial rather than transvaginal ultrasound, only 31 women had a placenta that was less than 3.0 cm from the internal os at 32 weeks of gestation.
Fig. 1. Prevalence of caesarean section in labouring women according to the distance between placental edge and internal cervical os. Means and their 95% confidence intervals are displayed. The dotted line represents the background caesarean section rate (23.8%).
The time interval between ultrasound and delivery has a significant influence on the classification and clinical outcome of placenta praevia. Several studies have demonstrated that the distance between the placental edge and the internal cervical os changes with advancing gestation6 – 13. Indeed, in one study, the mean rate of placental migration was estimated at 0.54 cm per week in the third trimester of pregnancy13. Thus, in the current study with an average of two weeks from ultrasound to delivery, the placenta is unlikely to have migrated more than 1.0 cm. In contrast, the placenta is likely to have migrated, on average, 2.7 and 3.2 cm in the two previous studies where the ultrasound to delivery interval was five and six weeks, respectively. The upper limit of the placental edge to internal os distance for the use of the term ‘praevia’ is unclear and undefined. Oppenheimer et al.3 included cases where this distance was up to 5.8 cm. Dawson et al.4 arbitrarily chose an upper limit of 3.0 cm. In the present study, an upper limit of 3.5 cm was used, as none of the cases of antepartum haemorrhage with a placental edge to internal os distance of over 3.5 cm were suspected to be due to placenta praevia. We acknowledge that this definition is arbitrary and based on clinical interpretation of ultrasound findings. However, analysis of the current data set allows the establishment of guidelines for delivery. Given the paucity of published evidence, the Royal College of Obstetricians and Gynaecologists14 recommend that trial of vaginal delivery is appropriate if the placental edge – internal os distance is over 2.0 cm. This guideline formed the basis for our post hoc definition of placenta praevia groups. The Royal College of Obstetricians and Gynaecologists’ recommendation is based on the findings of Oppenheimer et al.3, who performed a retrospective analysis of placenta praevia, where the data were grouped according to mode of delivery rather than antenatal ultrasound findings. They reported a mean placental edge to internal os distance measured five weeks prior to delivery of 1.1 cm (range 0 to 2.0 cm) in cases requiring a caesarean section (n ¼ 7), as opposed to 3.1 cm (range 1.8– 5.8 cm) in those delivering vaginally (n ¼ 14). The findings of the latter study are of limited clinical value because the grouping based on mode of delivery does not provide useful information for clinicians who need to make decisions based D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
PLACENTA PRAEVIA AND MODE OF DELIVERY
on antenatal ultrasound findings. The probability of vaginal delivery at a given placental edge – internal os distance cannot be calculated from the latter study. Indeed, the authors themselves concluded that, given the small numbers, the clinical significance of their findings was to be interpreted with caution and further studies were needed to establish guidelines. In the present study, the cohorts for comparison were identified on the basis of antenatal ultrasound findings. The findings of this study are that the likelihood for successful vaginal delivery increases with the distance of the placental edge from the internal os. This is clearly demonstrated in Table 3 and is illustrated in Fig. 1. Even though the emergency caesarean section rate was 90% in Group 2 and 37% in Group 3, these figures are likely to be influenced by clinical bias. Knowledge of the placental location, which was available to the clinicians managing the case, would probably have played an important part in the decision to perform an operative delivery. Hence, vaginal delivery rates in this study are likely to be conservative estimates. Clinicians may have been happier to conduct a trial of labour if they had sufficient scientific evidence to support an expectant management strategy. When evaluating the likelihood of vaginal delivery, it is clear from the data that placenta praevia reaching or overlapping the internal cervical os requires elective caesarean delivery. It is equally evident that when the placental edge is greater than 2 cm from the internal os, the likelihood of a successful trial of vaginal delivery is good. In Group 2, where the placental edge to internal os distance was 0.1 to 2 cm, some women achieved a successful vaginal delivery. However, all women with a placental edge to internal os distance of 1 cm or less required a caesarean delivery. Figure 1 also demonstrates a trend for decreasing caesarean section rates with increasing placental edge to internal os distances. Given the retrospective nature of the study and the likely clinical bias, management of cases with a placenta to internal os distance of 1.1 to 2 cm is unclear. A prospective study in women with placental edge – internal os distance of 1.1 to 2 cm may be needed to address this issue. It may be appropriate to allow these women a trial of labour with suitable precautions to manage emergency operative delivery and intrapartum/postpartum haemorrhage. The data of this study demonstrate that the likelihood for antepartum haemorrhage falls significantly from 57% in Group 1 to 28% in Group 3. It is important to note that a significant proportion (43%) of women with a placenta encroaching or overlapping the internal os did not have bleeding in the antenatal period. These data are supported with the two-in-three risk of antepartum haemorrhage reported in older series, where the diagnosis of placenta praevia was made on clinical grounds15. All the cases of postpartum haemorrhage in the present series occurred in women who had undergone a caesarean section. Five out of the 10 cases of postpartum haemorrhage took place in the group with the placenta overlapping D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864
863
the internal os. One of these five cases was a placenta accreta, with an estimated blood loss of 14 L, necessitating a hysterectomy. The postpartum haemorrhage rate in Groups 2 and 3 was similar to the background postpartum haemorrhage rate within our unit.
CONCLUSION The data of this study support the trial of vaginal delivery in cases with a placental to internal os distance >2 cm and an elective caesarean section when this distance is 1 cm or less. We propose that the term ‘praevia’ should be restricted to cases where the placental edge is 2 cm or less from the internal os, as the likelihood of operative delivery and significant postpartum haemorrhage remains high. Cases where the placenta is more than 2 cm from the internal os have a high chance of vaginal delivery and should be defined as ‘low lying’ rather than praevia, in order to reduce the bias towards operative delivery in these women.
Acknowledgements Dr Prefumo was supported by a Marie Curie Fellowship of the European Community programme Quality of Life under contract number QLGA-CT-2000-52145.
References 1. Conje JC, Taylor DJ. Bleeding in late pregnancy. In: James DK, Steer PJ, Weiner CP, Gonik B, editors. High Risk Pregnancy: Management Options, 2nd edition. Philadelphia: Saunders, 1999:111. 2. Oyelese Y. Placenta praevia and vasa praevia: time to leave the dark ages. Ultrasound Obstet Gynecol 2001;18:96 – 99. 3. Oppenheimer LW, Farine D, Knox Ritchie JW, Lewinsky RM, Telford J, Fairbanks LA. What is a low-lying placenta? Am J Obstet Gynecol 1991;165:1036 – 1038. 4. Dawson W, Dumas M, Romano W, Gagnon R, Gratton R, Mowbray D. Translabial ultrasonography and placenta previa: does measurement of the os – placenta distance predict outcome? J Ultrasound Med 1996; 15:441 – 446. 5. Obstetric hemorrhage. In: Cunningham FG, MacDonald PC, Gant NF, et al., editors. Williams Obstetrics, 20th edition. Stamford, Connecticut: Appleton and Lange, 1997:745 – 782. 6. Ruparelia BA, Chapman MG. Early low-lying placentae — ultrasonic assessment, progress and outcome. Eur J Obstet Gynecol Reprod Biol 1986;20:209 – 213. 7. Townsend RR, Laing FC, Nyberg DA, Brook JR, Wing VW. Technical factors responsible for ‘placental migration’: sonographic assessment. Radiology 1986;160:105 – 108. 8. Mustafa SA, Brizot ML, Carvalho MH, Watanabe L, Kanhale S, Zugaib M. Transvaginal ultrasonography in predicting placenta previa at delivery: a longitudinal study. Ultrasound Obstet Gynecol 2002;20: 356 – 359. 9. Lauria MR, Smith RS, Treadwelll MC, et al. The use of secondtrimester transvaginal sonography to predict placenta previa. Ultrasound Obstet Gynecol 1996;8:337 – 340. 10. Taipale P, Hiilesmaa V, Ylostalo P. Transvaginal ultrasonography at
864
A. BHIDE ET AL.
18 – 23 weeks in predicting placenta previa at delivery. Ultrasound Obstet Gynecol 1998;12:422 – 425. 11. Gallagher P, Fagan CJ, Bedi DG, Winsett MZ, Reyes RN. Potential placenta previa: definition, frequency, and significance. Am J Roentgenol 1987;149:1013 – 1015. 12. Dashe JS, McIntire DD, Ramus RM, Santos-Ramos R, Twickler DM. Persistence of placenta previa according to gestational age at ultrasound detection. Obstet Gynecol 2002;99:692 – 697. 13. Oppenheimer L, Simpson P, Holmes N, Dabrowski A. Diagnosis of
low-lying placenta: can migration in the third trimester predict outcome? Ultrasound Obstet Gynecol 2001;18:100 – 102. 14. Placenta Praevia: Diagnosis and Management. Clinical Green-top guideline no. 27. London: The Royal College of Obstetricians and Gynaecologists, 2001 (January). 15. Hill DJ, Beischer NA. Placenta praevia without anterpartum haemorrhage. Aust N Z J Obstet Gynaecol 1980;20:21 – 23. Accepted 28 May 2003
D RCOG 2003 Br J Obstet Gynaecol 110, pp. 860 – 864