- Email: [email protected]
Placenta previa: A 22-year analysis
Placenta previa: A 22-year analysis Marilynn C. Frederiksen, MD, Raymond Glassenberg, MD, and Catherine S. Stika, MD Chicago, Illinois OBJECTIVE: Our purpose was to identify what anesthetic method is safer for women with a placenta previa. STUDY DESIGN: We retrospectively reviewed all women with placenta previa who underwent cesarean delivery during the period January 1, 1976–December 31, 1997 at Northwestern Memorial Hospital. RESULTS: Of 93,384 deliveries, placenta previa was found in 514 women. Identifiable trends with time included an increasing incidence of placenta previa (r = 0.54, P < .01); cesarean hysterectomy (r = 0.54, P < .01); placenta accreta (r = 0.45, P < .03); and regional anesthesia (r = 0.84, P < .0001). The mean gestational age at delivery was 35.3 ± 3.4 weeks and did not change with time. General anesthesia was used for delivery in 380 women and regional anesthesia was used for 134 women. Prior cesarean delivery and general anesthesia were independent predictors of the need for blood transfusion, but only prior cesarean delivery was a predictor of the need for hysterectomy. General anesthesia increased the estimated blood loss, was associated with a lower postoperative hemoglobin concentration, and increased the need for blood transfusion. Elective and emergent deliveries did not differ in estimated blood loss, in postoperative hemoglobin concentrations, or in the incidence of intraoperative and anesthesia complications. Regional and general anesthesia did not differ in the incidence of intraoperative and anesthesia complications. CONCLUSIONS: In women with placenta previa, general anesthesia increased intraoperative blood loss and the need for blood transfusion. Regional anesthesia appears to be a safe alternative. (Am J Obstet Gynecol 1999;180:1432-7.)
Key words: Placenta previa, placenta accreta, cesarean delivery, general anesthesia, intraoperative blood loss
Although the management of a woman with placenta previa has become standardized, with expectant management favored until fetal lung maturity is established, followed by aggressively moving to delivery,1, 2 the type of anesthesia used for delivery is still controversial.3 The risk of excessive blood loss with delivery and the possibility of a cesarean hysterectomy in these patients prompts many anesthesiologists to prefer general anesthesia. There is evidence, however, to suggest that, in women with placenta previa, cesarean delivery with regional anesthesia decreases intraoperative blood loss in comparison with delivery under general anesthesia4-5 and that in selected women cesarean hysterectomy can be safely accomplished with regional anesthesia.6 For this study we hypothesized that general anesthesia is safer than regional anesthesia, particularly in women with a history of cesarean delivery and for any woman who might require a cesarean hysterectomy. Additionally, the study addressed From the Department of Obstetrics and Gynecology and the Department of Anesthesiology, Northwestern Memorial Hospital and Northwestern University Medical School. Poster Presentation, presented at the Sixty-sixth Annual Meeting of The Central Association of Obstetricians and Gynecologists, Kansas City, Missouri, October 15-17, 1998. Reprint requests: Marilynn C. Frederiksen, MD, 680 N Lake Shore Dr, Suite 1000, Chicago, IL 60611. Copyright © 1999 by Mosby, Inc. 0002-9378/99 $8.00 + 0 6/6/98245
1432
the following questions: (1) Is there a change in the incidence of placenta previa, placenta accreta, and the need for cesarean hysterectomy? (2) What are the risk factors for blood transfusion and for cesarean hysterectomy? (3) Is there a difference in the complications seen with elective cesarean delivery compared with emergent cesarean delivery in this population? Methods The hospital records of women delivered of their infants at Prentice Women’s Hospital of Northwestern Memorial Hospital during the period January 1, 1976–December 31, 1997 were searched to identify women who had placenta previa at the time of delivery. From 1976 to 1983 the hospital database alone was explored, using the discharge diagnosis of placenta previa to identify patients. Beginning in 1983 the perinatal database was used as a cross check with the hospital discharge diagnosis. Women with a low-lying placenta or a marginal placenta previa who were delivered vaginally were excluded. Medical records of identified women were abstracted for maternal age at delivery, gravidity, parity, gestational age at delivery, anesthesia type, preoperative and postoperative hemoglobin concentration, number of units of blood transfused, amount of crystalloid infused intraoperatively, intraoperative complications (related to both anesthesia and operative complications), and uter-
Volume 180, Number 6, Part 1 Am J Obstet Gynecol
Frederiksen, Glassenberg, and Stika 1433
Fig 1. Graph of incidence of cesarean delivery (squares with thick connecting line), number of patients with placenta previa (circles with broken connecting line), and number of cases of placenta accreta (squares with thin connecting line) during 22 years of study.
ine pathologic diagnosis if a cesarean hysterectomy was performed at the time of the delivery surgery. An elective cesarean delivery was defined as a surgical delivery performed after documentation of fetal lung maturity and without evidence of maternal hemorrhage. An emergent cesarean delivery was defined as a surgical delivery performed with evidence of recent maternal hemorrhage. The hemoglobin measurement just before surgery was used as the preoperative hemoglobin concentration, and the first hemoglobin measurement after surgery was used as the postoperative hemoglobin value. Anesthesia complications were defined as (1) maternal hypotension with a systolic arterial blood pressure of <80 mm Hg for ≥5 minutes, (2) evidence of hemodynamic instability as shown by a maternal arterial pH <7.2, or (3) disseminated intravascular coagulopathy defined by a partial thromboplastin time >50 seconds. The estimated blood loss used was from the anesthesia record. Operative complications were defined as unplanned surgery necessary either to repair damaged abdominal organs or to control bleeding or the need for a repeat laparotomy. Data were analyzed by means of descriptive statistical analysis, χ2 analysis, Fisher exact test, Student t test, linear regression, and Pearson correlation coefficient for bivariate analysis with the SPSS 7.1 software package (SPSS Inc, Chicago, Ill). Logistic regression was used to identify factors predictive of the need for blood transfusion and cesarean hysterectomy. For statistical tests, P < .05 was considered significant. Results From January 1, 1976–December 31, 1997, there were 93,384 deliveries and 514 women found to have placenta previa, for an incidence of placenta previa of 0.6%. The
Table I. Maternal characteristics of population with placenta previa Maternal characteristic Age (y)* Gravidity† 1 2 3 4 ≥5 Parity† Primiparous Multiparous Prior preterm birth† Prior abortion† Gestational age at delivery (wk)* Prior cesarean delivery† Maternal mortality rate
Women with placenta previa (n = 514) 31.4 ± 5.7 74 (14) 143 (28) 108 (21) 63 (12) 126 (25) 193 (38) 321 (62) 54 (11) 291 (57) 35.3 ± 3.4 106 (20) 0
*Values are expressed as mean ± SD. †Values are expressed as number and percent.
characteristics of the patient population are presented in Table I. The mean maternal age (±SD) of patients with placenta previa was 31.4 ± 5.7 years. The mean gestational age (±SD) at delivery was 35.3 ± 3.4 weeks. There were no maternal deaths in this population of women with placenta previa. Trends noted during the study period included an increase in the mean maternal age of women delivered with a diagnosis of placenta previa, rising from 29.9 ± 3.5 years in 1976 to 32.7 ± 5.5 years in 1997 (correlation coefficient, r = 0.81; P < .0001); an increase in the incidence of placenta previa with the annual incidence rising from 0.3% to 0.7% (correlation coefficient, r = 0.52, P < .01); an increase in the incidence of placenta accreta (correla-
1434 Frederiksen, Glassenberg, and Stika
June 1999 Am J Obstet Gynecol
Table II. Comparison of maternal characteristics between group delivered with general anesthesia and those delivered with regional anesthesia Characteristic Age (y)* Patients with prior cesarean delivery† Preoperative hemoglobin level (mg/dL)* Gestational age at delivery (wk)* Estimated blood loss (mL)* Intraoperative fluids (mL)* Transfusion required Postoperative hemoglobin level (mg/dL)*
General anesthesia (n = 380)
Regional anesthesia (n = 134)
Statistical significance
30.9 ± 5.9 91 (24) 11.4 ± 1.2 34.8 ± 3.6 1604 ± 1515 3923 ± 2836 115 (30) 8.8 ± 1.6
33.0 ± 4.8 15 (11) 12.0 ± 1.2 36.8 ± 2.2 1149 ± 848 3790 ± 1730 13 (9.7) 9.8 ± 1.4
P < .0001 P < .002 P < .0001 P < .0001 P < .001 NS P < .0001 P < .0001
NS, Not significant. *Values are expressed as mean ± SD. †Values are expressed as number and percent.
Table III. Characteristics of women requiring cesarean hysterectomy compared with women requiring cesarean delivery only Characteristic Age (y)* Patients with prior cesarean delivery† Preoperative hemoglobin level (mg/dL)* Fetal gestational age at delivery (wk)* Estimated blood loss (mL)* Intraoperative fluids (mL)* Transfusion required† Postoperative hemoglobin level (mg/dL)* Regional anesthesia†
Cesarean hysterectomy (n = 55) 32.9 ± 6.0 30 (55) 11.1 ± 1.1 34.5 ± 3.3 4060 ± 2440 8867 ± 4199 51 (93) 7.7 ± 1.5 7 (13)
Cesarean delivery (n = 459) 31.3 ± 5.6 76 (17) 11.6 ± 1.3 35.5 ± 3.4 1176 ± 727 3291 ± 1451 77 (17) 9.2 ± 1.5 127 (28)
Statistical significance NS P < .0001 NS NS P < .0001 P < .0001 P < .0001 P < .0001 P < .02
NS, Not significant. *Values are expressed as mean ± SD. †Values are expressed as number and percent.
tion coefficient, r = 0.45, P < .03); an increase in the need for cesarean hysterectomy (r = 0.54, P < .01); and an increase in the use of regional anesthesia (r = 0.84, P < .0001). During the study period the cesarean delivery rate of the institution varied from 16.4% to 27.2%, peaking in 1986 (Fig 1), and was not correlated with trends noted for the incidence of placenta previa, placenta accreta, or cesarean hysterectomy. The mean gestational age at delivery for women with a placenta previa did not change over time. Three hundred eighty women, or 74% of the population, were delivered while they were under general anesthesia. One hundred thirty-three women, or 26% of the population, were delivered while they were under regional anesthesia, and 1 woman had regional anesthesia that was later converted to general anesthesia. A comparison of the women delivered with general and those delivered with regional anesthesia is presented in Table II. Women who were delivered while they were under general anesthesia were younger than women delivered while they were under regional anesthesia, had a significantly higher incidence of prior cesarean delivery, a
lower preoperative hemoglobin concentration, and a shorter gestation at delivery. Women delivering under general anesthesia had a higher mean estimated blood loss, and lower postoperative hemoglobin than women delivering under regional anesthesia. Fluid replacement, however, was not significantly different between the 2 groups. Blood transfusions were required in only 25% of women. Transfusions were given to 115 of 380 women under general anesthesia or 30% and to 13 of 134 women under regional anesthesia or 9.7%, χ2 = 22.4, P < .0001. Factors identified as increasing the risk for a blood transfusion included general anesthesia, a history of a cesarean section, and the need to perform a cesarean hysterectomy. The odds ratio for a blood transfusion with delivery under general anesthesia was 3.1 (95% confidence interval, 1.85.3). Odds ratio for a blood transfusion with a history of a cesarean section was 3.0 (95% confidence interval, 1.94.8), and with the need for a hysterectomy the odds ratio rose to 63.3 (95% confidence interval, 22.2-180.2). For 408 women (79%), the delivery was their first cesarean delivery and 106 women (21%) had had at least 1 prior cesarean delivery (range, 1-4 prior cesarean deliver-
Frederiksen, Glassenberg, and Stika 1435
Volume 180, Number 6, Part 1 Am J Obstet Gynecol
Table IV. Characteristics of women requiring cesarean hysterectomy grouped according to whether surgery was done electively or on an emergency basis Maternal characteristic Age (y)* Gestational age at delivery (wk)* Patients with prior cesarean delivery† Preoperative hemoglobin level (mg/dL)* Estimated blood loss (mL)* Intraoperative fluids (mL)* Units transfused* Operative complications† Postoperative hemoglobin level (mg/dL)* Regional anesthesia† Anesthesia complications†
Elective cesarean delivery (n = 28)
Emergency cesarean delivery (n = 27)
Statistical significance
33.8 ± 5.2 36.6 ± 1.4 19 (68) 11.3 ± 1.2 4682 ± 2411 9968 ± 4253 6.4 ± 5.1 9 (32) 7.8 ± 1.5 6 (21) 9 (32)
31.7 ± 6.4 32.2 ± 3.2 11 (41) 11.0 ± 1.1 3452 ± 2491 7715 ± 3975 6.0 ± 5.0 4 (15) 7.6 ± 1.7 1 (4) 8 (30)
NS P < .0001 NS NS NS P < .05 NS NS NS NS NS
NS, Not significant. *Values are expressed as mean ± SD. †Values are expressed as number and percent.
ies). Fifty-five women required a hysterectomy at the time of delivery, for an incidence of cesarean hysterectomy of 11%. Pathologic changes in the hysterectomy specimens showed that 20 women (37%) had placenta accreta, 7 women (13%) had placenta increta, 7 women (13%) had placenta percreta), and 2 women (38%) were found to have no placental or uterine pathologic changes. The overall incidence of histologically confirmed placenta accreta in women with placenta previa was 34 (6.6%) of 514 women. In the population requiring a hysterectomy, the delivery was the first cesarean delivery for 25 women, whereas 30 women had a prior cesarean delivery. The need for cesarean hysterectomy occurred more commonly in women who had a prior cesarean delivery, 30 of 106 or 25%, as compared with women with no prior cesarean delivery, 25 of 408 or 6.1%, χ2 = 43.3, P < .0001. When we controlled for prior cesarean delivery and type of anesthesia, the need for cesarean hysterectomy was correlated with a history of cesarean delivery and not with the type of anesthesia. The odds ratio for cesarean hysterectomy in women with a history of a cesarean delivery compared with women without a prior cesarean delivery was 6.0 (95% confidence interval, 3.4-10.9). In a comparison of the group of women who required a cesarean hysterectomy with the group who did not (Table III), the women did not differ in mean maternal age, preoperative hemoglobin concentration, or mean duration of gestation at delivery. However, more women in the cesarean hysterectomy group than in the cesarean delivery (without hysterectomy) group had a prior cesarean delivery. Moreover, fewer women who required a cesarean hysterectomy were delivered under regional anesthesia. Grouping the women who required cesarean hysterectomy by whether the operation was elective or emergency (Table IV) showed only that the mean gestational age at delivery was significantly lower in the emergency group. The estimated blood loss, the blood units transfused, and
Table V. Operative and anesthetic complications that occurred with cesarean hysterectomy
Complication Intraoperative complication* Oophorectomy required Cystotomy Ureteral injury Bowel injury Hypogastric artery ligation Repeated laparotomy required* Anesthetic complication* Hemodynamic instability Disseminated intravascular coagulopathy Blood transfusions given intraoperatively Total units transfused†
Cesarean hysterectomy (n = 55)
1 (2) 6 (11) 1 (2) 0 6 (11) 5 (9) 10 (18) 12 (22) 51 6.2 ± 5.0
*Values are expressed as number and percent. †Values are expressed as mean ± SD.
the incidence of operative and anesthetic complications did not significantly differ in those women delivered electively and those women delivered on an emergency basis. The overall complications accompanying cesarean hysterectomy are presented in Table V. The most common complications were hemodynamic instability (18%) and disseminated intravascular coagulopathy (22%). The majority of women required intraoperative transfusion (mean number of units transfused, 6.2 ± 5.0). In a comparison of the anesthetic and operative complications encountered during cesarean hysterectomy with regional and general anesthesia, there was no significant difference between the 2 groups in anesthetic complications (1/7 vs 16/48; P value not significant), operative complications (0/7 vs 13/48; P value not significant), or required blood transfusions (6/7 vs 44/48; P value not significant). However, the mean number of units transfused was significantly less for the group having regional anesthesia (3.3 ± 1.8 units) than for the group having general anesthesia (6.6 ± 5.2 units; P < .005).
1436 Frederiksen, Glassenberg, and Stika
Comment Because our study period spanned 22 years, we first analyzed factors that changed during the study period. The increase in the mean age of women delivering with a placenta previa represents the change seen in the demographics of the women served by Northwestern Memorial Hospital and reflects a national trend in the general population where the mean age of first birth has risen and childbirth is delayed. Also identified is a doubling in the incidence of placenta previa and an increase in the incidence of placenta accreta and the need for cesarean hysterectomy. Although the mean incidence of placenta previa during our study period is similar to that found by Iyasu et al7 using the National Hospital Discharge Survey for years from 1979 through 1987, Iyasu et al found no increase in the incidence of placenta previa as we did, but our study extends over a longer period of time. Our institutional increase in the incidence of placenta previa may represent the effect of an increase in the rate of cesarean deliveries seen during this time period, not with a direct correlation but with a time lag to the occurrence of a placenta previa. Our study also confirmed the findings of others that prior cesarean section is an independent risk factor for cesarean hysterectomy and placenta accreta.8-10 The majority of the maternal morbidity from a placenta previa rests in those women who require a cesarean hysterectomy. In our previa population, despite similarities in mean maternal age, mean gestational age at delivery, and mean preoperative hemoglobin, the one identifiable risk factor for women who required a cesarean hysterectomy was a history of a cesarean section. Patients with a prior cesarean section had a 6.0 greater risk (95% confidence interval, 3.4-10.9) for a cesarean hysterectomy compared with patients without a prior cesarean section. As expected in women requiring a cesarean hysterectomy, we found higher estimated blood loss, a higher need for intraoperative fluids, a higher transfusion requirement, and a lower postoperative hemoglobin. To address the anesthesiologist’s concerns in managing women at high risk for anesthetic and operative complications and potentially a cesarean hysterectomy, we classified those patients who actually had a cesarean hysterectomy as to whether their delivery was emergent or elective, thinking that this mirrored the prospective clinical situation. In our population the women who required a cesarean hysterectomy were evenly dispersed between elective and emergent deliveries. No significant differences were found in the estimated blood loss, units transfused, or postoperative hemoglobins between the elective and emergent groups. Unlike Chestnut et al,6 who showed a significantly lower estimated blood loss, a lower need for intraoperative fluids, and a lower transfusion requirement in women undergoing cesarean hysterectomy as an elective procedure, we found no real dif-
June 1999 Am J Obstet Gynecol
ferences in these parameters in women requiring cesarean hysterectomy after an elective cesarean delivery as compared with women requiring a cesarean hysterectomy after an emergent cesarean delivery. The series of Chestnut et al,6 however, classified all hysterectomies with a placenta previa as emergent and included no category of elective cesarean delivery for placenta previa that then required hysterectomy. We were reassured to find no differences in the incidence of operative or anesthetic complications in women delivered electively or emergently who required hysterectomy. Although the numbers of cesarean hysterectomies performed with the patients under regional anesthesia in our series is small, it is also reassuring to find that there were no differences in anesthetic or operative complications encountered during cesarean hysterectomy with regional as compared with general anesthesia. However, the mean number of units of blood transfused was significantly more for the cesarean hysterectomy group under general anesthesia as compared with the group under regional anesthesia and is consistent with our findings in the larger population of all placenta previas. Therefore, having disproved our hypothesis, we have instead shown that a general anesthetic for women with a placenta previa increases the estimated blood loss, increases the risk for blood transfusion, and is associated with a lower postoperative hemoglobin. Risk factors for the need for blood transfusion include a prior cesarean section, the need for cesarean hysterectomy, and delivery with general anesthesia. There does not appear to be a difference in the complications seen with elective cesarean delivery as compared with emergent cesarean delivery in this population. Our findings support the use of regional anesthesia as a safe alternative to general anesthesia, even in women at high risk for a cesarean hysterectomy. Moreover, to further analyze factors beyond the limitation of a retrospective review, our study supports a randomized clinical trial comparing regional with general anesthesia for delivering women with a placenta previa in both emergent and elective situations. REFERENCES
1. Cotton DB, Read JA, Paul RH, Quilligan EJ. The conservative aggressive management of placenta previa. Am J Obstet Gynecol 1980;137:687-95. 2. Silver R, Deep R, Sabbagha RE, Dooley SL, Socol ML, Tamura RK. Placenta previa: aggressive expectant management. Am J Obstet Gynecol 1984;150:15-22. 3. Bonner SM, Haynes SR, Ryall D. The anaesthetic management of caesarean section for placenta praevia: a questionnaire survey. Anaesthesia 1995;50:992-4. 4. Arcario T, Greene M, Ostheimer GW, Datta S, Naulty JS. Risks of placenta previa/accreta in patients with previous cesarean deliveries [abstract]. Anesthesiology 1988;69:A659. 5. Abboud TK, Gerard C, Go A, Zhu J. Anesthesia for placenta previa at Women’s Hospital: a 3 year survey. In: Proceedings of the Twenty-fifth Annual Meeting of the SOAP; 1993 May. p. 19. 6. Chestnut DH, Dewan DM, Redick LF, Caton D, Spielman FJ.
Volume 180, Number 6, Part 1 Am J Obstet Gynecol
Anesthetic management for obstetric hysterectomy: a multi-institutional study. Anesthesiology 1989;70:607-10. 7. Iyasu S, Saftlas AK, Rowley DL, Koonin LM, Lawson HW, Atrash HK. The epidemiology of placenta previa in the United States, 1979 through 1987. Am J Obstet Gynecol 1993;168:1424-9. 8. Clark SL, Koonings PP, Phelan JP. Placenta previa/accreta and prior cesarean section. Obstet Gynecol 1985;66:89-91
Frederiksen, Glassenberg, and Stika 1437
9. Clark SL, Yeh SY, Phelan J-P, Bruce S, Paul RH. Emergency hysterectomy for obstetric hemorrhage. Obstet Gynecol 1984;64:376-80. 10. Singh PM, Rodriques C, Gupta AN. Placenta previa and previous cesarean section. Acta Obstet Gynecol Scand 1981;60: 367-8. 11. Dinsmoor MJ, Hogg BB. Autologous blood donation with placenta previa: is it feasible? Am J Perinatol 1995;12:382-4.
Bound volumes available to subscribers Bound volumes of the American Journal of Obstetrics and Gynecology are available to subscribers (only) for the 1999 issues from the publisher, at a cost of $122.00 for domestic, $156.22 for Canada, and $146.00 for international for Vol. 180 (January-June) and Vol. 181 (July-December). Shipping charges are included. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Inc, Subscription Services, 11830 Westline Industrial Dr, St Louis, MO 63146-3318, USA; telephone (800)453-4351 or (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.
Key words: Placenta previa, placenta accreta, cesarean delivery, general anesthesia, intraoperative blood loss
Although the management of a woman with placenta previa has become standardized, with expectant management favored until fetal lung maturity is established, followed by aggressively moving to delivery,1, 2 the type of anesthesia used for delivery is still controversial.3 The risk of excessive blood loss with delivery and the possibility of a cesarean hysterectomy in these patients prompts many anesthesiologists to prefer general anesthesia. There is evidence, however, to suggest that, in women with placenta previa, cesarean delivery with regional anesthesia decreases intraoperative blood loss in comparison with delivery under general anesthesia4-5 and that in selected women cesarean hysterectomy can be safely accomplished with regional anesthesia.6 For this study we hypothesized that general anesthesia is safer than regional anesthesia, particularly in women with a history of cesarean delivery and for any woman who might require a cesarean hysterectomy. Additionally, the study addressed From the Department of Obstetrics and Gynecology and the Department of Anesthesiology, Northwestern Memorial Hospital and Northwestern University Medical School. Poster Presentation, presented at the Sixty-sixth Annual Meeting of The Central Association of Obstetricians and Gynecologists, Kansas City, Missouri, October 15-17, 1998. Reprint requests: Marilynn C. Frederiksen, MD, 680 N Lake Shore Dr, Suite 1000, Chicago, IL 60611. Copyright © 1999 by Mosby, Inc. 0002-9378/99 $8.00 + 0 6/6/98245
1432
the following questions: (1) Is there a change in the incidence of placenta previa, placenta accreta, and the need for cesarean hysterectomy? (2) What are the risk factors for blood transfusion and for cesarean hysterectomy? (3) Is there a difference in the complications seen with elective cesarean delivery compared with emergent cesarean delivery in this population? Methods The hospital records of women delivered of their infants at Prentice Women’s Hospital of Northwestern Memorial Hospital during the period January 1, 1976–December 31, 1997 were searched to identify women who had placenta previa at the time of delivery. From 1976 to 1983 the hospital database alone was explored, using the discharge diagnosis of placenta previa to identify patients. Beginning in 1983 the perinatal database was used as a cross check with the hospital discharge diagnosis. Women with a low-lying placenta or a marginal placenta previa who were delivered vaginally were excluded. Medical records of identified women were abstracted for maternal age at delivery, gravidity, parity, gestational age at delivery, anesthesia type, preoperative and postoperative hemoglobin concentration, number of units of blood transfused, amount of crystalloid infused intraoperatively, intraoperative complications (related to both anesthesia and operative complications), and uter-
Volume 180, Number 6, Part 1 Am J Obstet Gynecol
Frederiksen, Glassenberg, and Stika 1433
Fig 1. Graph of incidence of cesarean delivery (squares with thick connecting line), number of patients with placenta previa (circles with broken connecting line), and number of cases of placenta accreta (squares with thin connecting line) during 22 years of study.
ine pathologic diagnosis if a cesarean hysterectomy was performed at the time of the delivery surgery. An elective cesarean delivery was defined as a surgical delivery performed after documentation of fetal lung maturity and without evidence of maternal hemorrhage. An emergent cesarean delivery was defined as a surgical delivery performed with evidence of recent maternal hemorrhage. The hemoglobin measurement just before surgery was used as the preoperative hemoglobin concentration, and the first hemoglobin measurement after surgery was used as the postoperative hemoglobin value. Anesthesia complications were defined as (1) maternal hypotension with a systolic arterial blood pressure of <80 mm Hg for ≥5 minutes, (2) evidence of hemodynamic instability as shown by a maternal arterial pH <7.2, or (3) disseminated intravascular coagulopathy defined by a partial thromboplastin time >50 seconds. The estimated blood loss used was from the anesthesia record. Operative complications were defined as unplanned surgery necessary either to repair damaged abdominal organs or to control bleeding or the need for a repeat laparotomy. Data were analyzed by means of descriptive statistical analysis, χ2 analysis, Fisher exact test, Student t test, linear regression, and Pearson correlation coefficient for bivariate analysis with the SPSS 7.1 software package (SPSS Inc, Chicago, Ill). Logistic regression was used to identify factors predictive of the need for blood transfusion and cesarean hysterectomy. For statistical tests, P < .05 was considered significant. Results From January 1, 1976–December 31, 1997, there were 93,384 deliveries and 514 women found to have placenta previa, for an incidence of placenta previa of 0.6%. The
Table I. Maternal characteristics of population with placenta previa Maternal characteristic Age (y)* Gravidity† 1 2 3 4 ≥5 Parity† Primiparous Multiparous Prior preterm birth† Prior abortion† Gestational age at delivery (wk)* Prior cesarean delivery† Maternal mortality rate
Women with placenta previa (n = 514) 31.4 ± 5.7 74 (14) 143 (28) 108 (21) 63 (12) 126 (25) 193 (38) 321 (62) 54 (11) 291 (57) 35.3 ± 3.4 106 (20) 0
*Values are expressed as mean ± SD. †Values are expressed as number and percent.
characteristics of the patient population are presented in Table I. The mean maternal age (±SD) of patients with placenta previa was 31.4 ± 5.7 years. The mean gestational age (±SD) at delivery was 35.3 ± 3.4 weeks. There were no maternal deaths in this population of women with placenta previa. Trends noted during the study period included an increase in the mean maternal age of women delivered with a diagnosis of placenta previa, rising from 29.9 ± 3.5 years in 1976 to 32.7 ± 5.5 years in 1997 (correlation coefficient, r = 0.81; P < .0001); an increase in the incidence of placenta previa with the annual incidence rising from 0.3% to 0.7% (correlation coefficient, r = 0.52, P < .01); an increase in the incidence of placenta accreta (correla-
1434 Frederiksen, Glassenberg, and Stika
June 1999 Am J Obstet Gynecol
Table II. Comparison of maternal characteristics between group delivered with general anesthesia and those delivered with regional anesthesia Characteristic Age (y)* Patients with prior cesarean delivery† Preoperative hemoglobin level (mg/dL)* Gestational age at delivery (wk)* Estimated blood loss (mL)* Intraoperative fluids (mL)* Transfusion required Postoperative hemoglobin level (mg/dL)*
General anesthesia (n = 380)
Regional anesthesia (n = 134)
Statistical significance
30.9 ± 5.9 91 (24) 11.4 ± 1.2 34.8 ± 3.6 1604 ± 1515 3923 ± 2836 115 (30) 8.8 ± 1.6
33.0 ± 4.8 15 (11) 12.0 ± 1.2 36.8 ± 2.2 1149 ± 848 3790 ± 1730 13 (9.7) 9.8 ± 1.4
P < .0001 P < .002 P < .0001 P < .0001 P < .001 NS P < .0001 P < .0001
NS, Not significant. *Values are expressed as mean ± SD. †Values are expressed as number and percent.
Table III. Characteristics of women requiring cesarean hysterectomy compared with women requiring cesarean delivery only Characteristic Age (y)* Patients with prior cesarean delivery† Preoperative hemoglobin level (mg/dL)* Fetal gestational age at delivery (wk)* Estimated blood loss (mL)* Intraoperative fluids (mL)* Transfusion required† Postoperative hemoglobin level (mg/dL)* Regional anesthesia†
Cesarean hysterectomy (n = 55) 32.9 ± 6.0 30 (55) 11.1 ± 1.1 34.5 ± 3.3 4060 ± 2440 8867 ± 4199 51 (93) 7.7 ± 1.5 7 (13)
Cesarean delivery (n = 459) 31.3 ± 5.6 76 (17) 11.6 ± 1.3 35.5 ± 3.4 1176 ± 727 3291 ± 1451 77 (17) 9.2 ± 1.5 127 (28)
Statistical significance NS P < .0001 NS NS P < .0001 P < .0001 P < .0001 P < .0001 P < .02
NS, Not significant. *Values are expressed as mean ± SD. †Values are expressed as number and percent.
tion coefficient, r = 0.45, P < .03); an increase in the need for cesarean hysterectomy (r = 0.54, P < .01); and an increase in the use of regional anesthesia (r = 0.84, P < .0001). During the study period the cesarean delivery rate of the institution varied from 16.4% to 27.2%, peaking in 1986 (Fig 1), and was not correlated with trends noted for the incidence of placenta previa, placenta accreta, or cesarean hysterectomy. The mean gestational age at delivery for women with a placenta previa did not change over time. Three hundred eighty women, or 74% of the population, were delivered while they were under general anesthesia. One hundred thirty-three women, or 26% of the population, were delivered while they were under regional anesthesia, and 1 woman had regional anesthesia that was later converted to general anesthesia. A comparison of the women delivered with general and those delivered with regional anesthesia is presented in Table II. Women who were delivered while they were under general anesthesia were younger than women delivered while they were under regional anesthesia, had a significantly higher incidence of prior cesarean delivery, a
lower preoperative hemoglobin concentration, and a shorter gestation at delivery. Women delivering under general anesthesia had a higher mean estimated blood loss, and lower postoperative hemoglobin than women delivering under regional anesthesia. Fluid replacement, however, was not significantly different between the 2 groups. Blood transfusions were required in only 25% of women. Transfusions were given to 115 of 380 women under general anesthesia or 30% and to 13 of 134 women under regional anesthesia or 9.7%, χ2 = 22.4, P < .0001. Factors identified as increasing the risk for a blood transfusion included general anesthesia, a history of a cesarean section, and the need to perform a cesarean hysterectomy. The odds ratio for a blood transfusion with delivery under general anesthesia was 3.1 (95% confidence interval, 1.85.3). Odds ratio for a blood transfusion with a history of a cesarean section was 3.0 (95% confidence interval, 1.94.8), and with the need for a hysterectomy the odds ratio rose to 63.3 (95% confidence interval, 22.2-180.2). For 408 women (79%), the delivery was their first cesarean delivery and 106 women (21%) had had at least 1 prior cesarean delivery (range, 1-4 prior cesarean deliver-
Frederiksen, Glassenberg, and Stika 1435
Volume 180, Number 6, Part 1 Am J Obstet Gynecol
Table IV. Characteristics of women requiring cesarean hysterectomy grouped according to whether surgery was done electively or on an emergency basis Maternal characteristic Age (y)* Gestational age at delivery (wk)* Patients with prior cesarean delivery† Preoperative hemoglobin level (mg/dL)* Estimated blood loss (mL)* Intraoperative fluids (mL)* Units transfused* Operative complications† Postoperative hemoglobin level (mg/dL)* Regional anesthesia† Anesthesia complications†
Elective cesarean delivery (n = 28)
Emergency cesarean delivery (n = 27)
Statistical significance
33.8 ± 5.2 36.6 ± 1.4 19 (68) 11.3 ± 1.2 4682 ± 2411 9968 ± 4253 6.4 ± 5.1 9 (32) 7.8 ± 1.5 6 (21) 9 (32)
31.7 ± 6.4 32.2 ± 3.2 11 (41) 11.0 ± 1.1 3452 ± 2491 7715 ± 3975 6.0 ± 5.0 4 (15) 7.6 ± 1.7 1 (4) 8 (30)
NS P < .0001 NS NS NS P < .05 NS NS NS NS NS
NS, Not significant. *Values are expressed as mean ± SD. †Values are expressed as number and percent.
ies). Fifty-five women required a hysterectomy at the time of delivery, for an incidence of cesarean hysterectomy of 11%. Pathologic changes in the hysterectomy specimens showed that 20 women (37%) had placenta accreta, 7 women (13%) had placenta increta, 7 women (13%) had placenta percreta), and 2 women (38%) were found to have no placental or uterine pathologic changes. The overall incidence of histologically confirmed placenta accreta in women with placenta previa was 34 (6.6%) of 514 women. In the population requiring a hysterectomy, the delivery was the first cesarean delivery for 25 women, whereas 30 women had a prior cesarean delivery. The need for cesarean hysterectomy occurred more commonly in women who had a prior cesarean delivery, 30 of 106 or 25%, as compared with women with no prior cesarean delivery, 25 of 408 or 6.1%, χ2 = 43.3, P < .0001. When we controlled for prior cesarean delivery and type of anesthesia, the need for cesarean hysterectomy was correlated with a history of cesarean delivery and not with the type of anesthesia. The odds ratio for cesarean hysterectomy in women with a history of a cesarean delivery compared with women without a prior cesarean delivery was 6.0 (95% confidence interval, 3.4-10.9). In a comparison of the group of women who required a cesarean hysterectomy with the group who did not (Table III), the women did not differ in mean maternal age, preoperative hemoglobin concentration, or mean duration of gestation at delivery. However, more women in the cesarean hysterectomy group than in the cesarean delivery (without hysterectomy) group had a prior cesarean delivery. Moreover, fewer women who required a cesarean hysterectomy were delivered under regional anesthesia. Grouping the women who required cesarean hysterectomy by whether the operation was elective or emergency (Table IV) showed only that the mean gestational age at delivery was significantly lower in the emergency group. The estimated blood loss, the blood units transfused, and
Table V. Operative and anesthetic complications that occurred with cesarean hysterectomy
Complication Intraoperative complication* Oophorectomy required Cystotomy Ureteral injury Bowel injury Hypogastric artery ligation Repeated laparotomy required* Anesthetic complication* Hemodynamic instability Disseminated intravascular coagulopathy Blood transfusions given intraoperatively Total units transfused†
Cesarean hysterectomy (n = 55)
1 (2) 6 (11) 1 (2) 0 6 (11) 5 (9) 10 (18) 12 (22) 51 6.2 ± 5.0
*Values are expressed as number and percent. †Values are expressed as mean ± SD.
the incidence of operative and anesthetic complications did not significantly differ in those women delivered electively and those women delivered on an emergency basis. The overall complications accompanying cesarean hysterectomy are presented in Table V. The most common complications were hemodynamic instability (18%) and disseminated intravascular coagulopathy (22%). The majority of women required intraoperative transfusion (mean number of units transfused, 6.2 ± 5.0). In a comparison of the anesthetic and operative complications encountered during cesarean hysterectomy with regional and general anesthesia, there was no significant difference between the 2 groups in anesthetic complications (1/7 vs 16/48; P value not significant), operative complications (0/7 vs 13/48; P value not significant), or required blood transfusions (6/7 vs 44/48; P value not significant). However, the mean number of units transfused was significantly less for the group having regional anesthesia (3.3 ± 1.8 units) than for the group having general anesthesia (6.6 ± 5.2 units; P < .005).
1436 Frederiksen, Glassenberg, and Stika
Comment Because our study period spanned 22 years, we first analyzed factors that changed during the study period. The increase in the mean age of women delivering with a placenta previa represents the change seen in the demographics of the women served by Northwestern Memorial Hospital and reflects a national trend in the general population where the mean age of first birth has risen and childbirth is delayed. Also identified is a doubling in the incidence of placenta previa and an increase in the incidence of placenta accreta and the need for cesarean hysterectomy. Although the mean incidence of placenta previa during our study period is similar to that found by Iyasu et al7 using the National Hospital Discharge Survey for years from 1979 through 1987, Iyasu et al found no increase in the incidence of placenta previa as we did, but our study extends over a longer period of time. Our institutional increase in the incidence of placenta previa may represent the effect of an increase in the rate of cesarean deliveries seen during this time period, not with a direct correlation but with a time lag to the occurrence of a placenta previa. Our study also confirmed the findings of others that prior cesarean section is an independent risk factor for cesarean hysterectomy and placenta accreta.8-10 The majority of the maternal morbidity from a placenta previa rests in those women who require a cesarean hysterectomy. In our previa population, despite similarities in mean maternal age, mean gestational age at delivery, and mean preoperative hemoglobin, the one identifiable risk factor for women who required a cesarean hysterectomy was a history of a cesarean section. Patients with a prior cesarean section had a 6.0 greater risk (95% confidence interval, 3.4-10.9) for a cesarean hysterectomy compared with patients without a prior cesarean section. As expected in women requiring a cesarean hysterectomy, we found higher estimated blood loss, a higher need for intraoperative fluids, a higher transfusion requirement, and a lower postoperative hemoglobin. To address the anesthesiologist’s concerns in managing women at high risk for anesthetic and operative complications and potentially a cesarean hysterectomy, we classified those patients who actually had a cesarean hysterectomy as to whether their delivery was emergent or elective, thinking that this mirrored the prospective clinical situation. In our population the women who required a cesarean hysterectomy were evenly dispersed between elective and emergent deliveries. No significant differences were found in the estimated blood loss, units transfused, or postoperative hemoglobins between the elective and emergent groups. Unlike Chestnut et al,6 who showed a significantly lower estimated blood loss, a lower need for intraoperative fluids, and a lower transfusion requirement in women undergoing cesarean hysterectomy as an elective procedure, we found no real dif-
June 1999 Am J Obstet Gynecol
ferences in these parameters in women requiring cesarean hysterectomy after an elective cesarean delivery as compared with women requiring a cesarean hysterectomy after an emergent cesarean delivery. The series of Chestnut et al,6 however, classified all hysterectomies with a placenta previa as emergent and included no category of elective cesarean delivery for placenta previa that then required hysterectomy. We were reassured to find no differences in the incidence of operative or anesthetic complications in women delivered electively or emergently who required hysterectomy. Although the numbers of cesarean hysterectomies performed with the patients under regional anesthesia in our series is small, it is also reassuring to find that there were no differences in anesthetic or operative complications encountered during cesarean hysterectomy with regional as compared with general anesthesia. However, the mean number of units of blood transfused was significantly more for the cesarean hysterectomy group under general anesthesia as compared with the group under regional anesthesia and is consistent with our findings in the larger population of all placenta previas. Therefore, having disproved our hypothesis, we have instead shown that a general anesthetic for women with a placenta previa increases the estimated blood loss, increases the risk for blood transfusion, and is associated with a lower postoperative hemoglobin. Risk factors for the need for blood transfusion include a prior cesarean section, the need for cesarean hysterectomy, and delivery with general anesthesia. There does not appear to be a difference in the complications seen with elective cesarean delivery as compared with emergent cesarean delivery in this population. Our findings support the use of regional anesthesia as a safe alternative to general anesthesia, even in women at high risk for a cesarean hysterectomy. Moreover, to further analyze factors beyond the limitation of a retrospective review, our study supports a randomized clinical trial comparing regional with general anesthesia for delivering women with a placenta previa in both emergent and elective situations. REFERENCES
1. Cotton DB, Read JA, Paul RH, Quilligan EJ. The conservative aggressive management of placenta previa. Am J Obstet Gynecol 1980;137:687-95. 2. Silver R, Deep R, Sabbagha RE, Dooley SL, Socol ML, Tamura RK. Placenta previa: aggressive expectant management. Am J Obstet Gynecol 1984;150:15-22. 3. Bonner SM, Haynes SR, Ryall D. The anaesthetic management of caesarean section for placenta praevia: a questionnaire survey. Anaesthesia 1995;50:992-4. 4. Arcario T, Greene M, Ostheimer GW, Datta S, Naulty JS. Risks of placenta previa/accreta in patients with previous cesarean deliveries [abstract]. Anesthesiology 1988;69:A659. 5. Abboud TK, Gerard C, Go A, Zhu J. Anesthesia for placenta previa at Women’s Hospital: a 3 year survey. In: Proceedings of the Twenty-fifth Annual Meeting of the SOAP; 1993 May. p. 19. 6. Chestnut DH, Dewan DM, Redick LF, Caton D, Spielman FJ.
Volume 180, Number 6, Part 1 Am J Obstet Gynecol
Anesthetic management for obstetric hysterectomy: a multi-institutional study. Anesthesiology 1989;70:607-10. 7. Iyasu S, Saftlas AK, Rowley DL, Koonin LM, Lawson HW, Atrash HK. The epidemiology of placenta previa in the United States, 1979 through 1987. Am J Obstet Gynecol 1993;168:1424-9. 8. Clark SL, Koonings PP, Phelan JP. Placenta previa/accreta and prior cesarean section. Obstet Gynecol 1985;66:89-91
Frederiksen, Glassenberg, and Stika 1437
9. Clark SL, Yeh SY, Phelan J-P, Bruce S, Paul RH. Emergency hysterectomy for obstetric hemorrhage. Obstet Gynecol 1984;64:376-80. 10. Singh PM, Rodriques C, Gupta AN. Placenta previa and previous cesarean section. Acta Obstet Gynecol Scand 1981;60: 367-8. 11. Dinsmoor MJ, Hogg BB. Autologous blood donation with placenta previa: is it feasible? Am J Perinatol 1995;12:382-4.
Bound volumes available to subscribers Bound volumes of the American Journal of Obstetrics and Gynecology are available to subscribers (only) for the 1999 issues from the publisher, at a cost of $122.00 for domestic, $156.22 for Canada, and $146.00 for international for Vol. 180 (January-June) and Vol. 181 (July-December). Shipping charges are included. Each bound volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the Journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Mosby, Inc, Subscription Services, 11830 Westline Industrial Dr, St Louis, MO 63146-3318, USA; telephone (800)453-4351 or (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular Journal subscription.