Clinical risk factors for placenta previa-placenta accreta

David A. Miller, MD, Janet A. Chollet, MD, and T. Murphy Goodwin, MD

Los Angeles, California
OBJECTIVE: Our purpose was to define the clinical risk factors associated with placenta previa-placenta accreta. STUDY DESIGN: Hospital records were reviewed of all cases of placenta accreta confirmed histologically between January 1985 and December 1994. Additionally, we reviewed the records of all women with placenta previa and all those undergoing cesarean hysterectomy during the same period. Multiple logistic regression analysis was used to identify independent clinical risk factors for placenta accreta. RESULTS: Among 155,670 deliveries, 62 (1/2510) were complicated by histologically confirmed placenta accreta. Placenta accreta occurred in 55 of 590 (9.3%) women with placenta previa and in 7 of 155,080 (1/22,154) without placenta previa (relative risk 2065, 95% confidence interval 944 to 4516, p < 0.0001). Among women with placenta previa, advanced maternal age (->35 years) and previous cesarean delivery were independent risk factors for placenta accreta, Placenta accreta was present in 36 of 124 (29%) cases in which the placenta was implanted over the uterine scar and in 4 of 62 (6.5%) cases in which it was not (relatiave risk 4.5, 95% confidence interval 1.68 to 12.07). Among women with placenta previa, the risk of placenta accreta ranged from 2% in women <35 years old with no previous cesarean deliveries to almost 39% in women with two or more previous cesarean deliveries and an anterior or central placenta previa. CONCLUSION: Placenta accreta occurs in approximately 1 of 2500 deliveries. Among women with placenta previa, the incidence is nearly 10%. In this high-risk group advanced maternal age and previous cesarean section are independent risk factors. (Am J Obstet Gynecol 1997;177:210-4.)

Key words: Placenta previa, placenta accreta, incidence, risk factors

Placenta accreta is an abnormally firm attachment of placental villi to the uterine wall with absence of the normal intervening decidua basalis and fibrinoid layer of Nitabuch. Three variants of the condition are generally recognized. In the most common form, placenta accreta, the placenta is attached directly to the myometrium. Less commonly, it may extend into the myometrium (placenta increta) or through the entire myometrial layer ( placenta percreta). The reported incidence varies widely, from 1 in 540 to 1 in 70,000 deliveries, with an average incidence of approximately 1 in 7000.1 The incidence is highest among women with placenta previa or a previous cesarean section. Other purported risk factors include advanced maternal age, mulfiparity, and previous uterine curettage. We performed a retrospective analysis of 62 histopathologically confirmed cases of placenta accreta in an attempt to

identify i n d e p e n d e n t risk factors and to determine the magnitude of risk conferred by each.
Material and methods

Hospital records were reviewed of all cases of placenta accreta confirmed by histologic examination of hysterectomy specimens between Jan. 1, 1985, and Dec. 31, 1994. Additionally, we reviewed the records of all women undergoing cesarean hysterectomy and all women with placenta previa during the same period. Multiple logistic regression analysis was used to identify i n d e p e n d e n t risk factors for placenta accreta. Relative risks were calculated with Taylor series 95% confidence intervals. A p value of 0.05 was considered statistically significant.
Results

From the Division of MaternaI-Fetal Medicine, Department of Obstetrics and Gynecology, Los Angeles' County~University of Southern California Women's and Children's Hospital. Receivedfor publication September 16, 1996; revisedFebruary 5, 1997; acceptedFebruary 20, 1997. Reprint requests: David A. Miller, MD, LAC+USC Women's and Children's Hospital, Department of Obstetrics and Gynecology, Room 5K-40, 1240 North Mission Road, Los Angeles, CA 90033. Copyright 1997 by Mosby-Year Book, Inc. 0002-9378/97 $5.00 + 0 6/1/81387

During the study period, there were 155,670 deliveries at Los Angeles County/University of Southern California Women's Hospital. One h u n d r e d twenty-seven (1/1226) cesarean hysterectomies were performed with an operative diagnosis of placenta accreta. In 65 cases (52%) histologic findings did not confirm the diagnosis of placenta accreta. The remaining 62 cases (48%) were confirmed histologically and are the subjects of our report. Among these, 47 were placenta accreta, 11 were placenta increta, and 4 were placenta percreta. The

210

Volume 177, Number ]

AmJ Obstet Gynecol

Miller, Chollet, and Goodwin

211

Table I. Relationship of placenta previa to n u m b e r of previous cesarean sections

Previous cesarean sections (No.) Incidence of placenta previa* (%)

Table II. I n d e p e n d e n t risk factors for placenta accreta among women with placenta previa
95% confidence Relative risk interval Significance

Risk factor

0 1 2 ->3

0.3 0.8 2.0 4.2

*All differences statistically significant (p < 0.01).

Advanced maternal age (->35 yr) Previous cesarean section 1 >2

2.28

1.25-4.16 p < 0.01

4.45 11.32

2.00-0.50 p < 0.0001 5.50-22.02 p < 0.0001

overall incidence of confirmed placenta accreta was 1 in 2510. The mean age of women with placenta accreta was 33 years, gravidity 5.2, parity 3.4, n u m b e r of previous cesarean deliveries 1.6, and n u m b e r of previous curettage procedures 0.8. Fifty-five (89%) women with placenta accreta had coexisting placenta previa, 45 (73%) had at least one previous cesarean section, and 30 (48%) had at least one previous curettage. Forty-one (66%) were multiparous (parity-->3), 11 were grandmultiparous (parity ->5), and 28 (45%) were ->35 years old. There were 590 (1/263) cases of placenta previa during the same period. The incidence of placenta previa was directly related to the n u m b e r of previous cesarean sections (Table I). Two h u n d r e d ninety women with placenta previa (50%) had at least one previous curettage, and 144 (24%) had two or more. Among 590 women with placenta previa, 55 (9.3%) had placenta accreta. By comparison, placenta accreta occurred in 7 of 155,080 (0.005%) women without placenta previa (relative risk 2065, 95% confidence interval 944 to 4516, p < 0.0001). Placenta accreta occurred in 44 of 17,424 women (0.25%) with at least one previous cesarean delivery and in 16 of 138,246 (0.01%) without a previous cesarean (relative risk 19.39, 95% confidence interval 11.21 to 33.55, p < 0.0001). Among 186 women with placenta previa and a previous cesarean delivery, the incidence of placenta accreta was 22%; with neither placenta previa nor a previous cesarean section, the incidence was 1 in 68,000. Because the absence of placenta previa virtually excluded the diagnosis of placenta accreta and its presence conferred a risk of nearly 10%, analysis of additional risk factors was confined to the high-risk group of women with placenta previa. Maternal age -->35years and previous cesarean section were identified as i n d e p e n d e n t risk factors for placenta accreta (Table II). Placenta accreta was present in 3.2% of women younger than 25 years, 6.2% of those aged 25 to 29 years, 10.2% of women 30 to 34 years old, and 14.6% beyond age 34 years. Similarly, placenta accreta was present in 4% of women with placenta previa and no previous cesarean deliveries, 14% with one previous cesarean section, 23% with two, 35%

with three, and 50% with four previous cesarean deliveries. Multiparity and grandmultiparity were significantly associated with placenta accreta, as well. However, statistical significance did not persist after we corrected for advanced maternal age. Among women with placenta previa, previous uterine curettage was not a significant risk factor for placenta accreta. In women with placenta previa and at least one previous cesarean section, the risk of placenta accreta was further modified by the location of the placenta with respect to the uterine scar. Placenta accreta occurred in 36 of 124 (29%) cases in which the placenta was implanted anteriorly or centrally, overlying the uterine scar, and in 4 of 62 (6.5%) cases in which it was not (relative risk 4.5, 95% confidence interval 1.68 to 12.07). The parameters of maternal age, n u m b e r of previous cesarean deliveries, and placental location with respect to the uterine scar were combined to generate differential estimates of the risk of placenta accreta in the presence of placenta previa (Table III). Prematurity was the primary neonatal complication associated with placenta accreta. Among 62 cases of placenta accreta the mean gestational age at delivery was 34.6 weeks (+5.9). Gestational age was <37 weeks in 37 (59%) cases, <34 weeks in 16 (26%), and <30 weeks in 3 (5%). The mean birth weight was 2791 gm (_+923). Birth weight was --<2500 gm in 22 (35%) and -<1500 gm in 5 (8%). Three neonates (5%) weighed <10th percentile for gestational age. 2 Five-minute Apgar scores were < 7 in three cases, and there were no neonatal deaths. The mean gestational age at delivery in 55 cases of placenta previa-placenta accreta (34.9 _+ 4.0 weeks) was similar to that in 535 cases of placenta previa without placenta accreta (34.9 _+ 3.9 weeks, p = 0.99). Among women with placenta previa-placenta accreta the gestational age at delivery was <37 weeks in 36 (65%), <34 weeks in 17 (30.9%), and <30 weeks in 4 (7%). Among women with placenta previa without placenta accreta the gestational age at delivery was <37 weeks in 337 (63%), <34 weeks in 171 (32%), and <30 weeks in 76 (14%). Among 7 cases of placenta accreta without placenta

212

Miller, Chollet, and Goodwin

july 1997 AmJ Obstet Gynecol

Table III. Estimated risk of placenta accreta among women with placenta previa.
Placenta not overlying uterine scar Previous cesarean sections Age < 3 5 yr Age >--35 yr Placenta overlying uterine scar Age < 3 5 yr Age > 3 5 yr

0 1 >2
N/A, Not applicable.

2.1% (6/288) 3.7% (1/27) 5.2% (1/19)

6.3% (9/144) 9.1% (1/!1) 20% (1/5)

N/A 15.9% (7/44) 38.5% (15/39)

N/A 30% (6/20) 38.1% (8/21)

previa the mean gestational age at delivery was 37 weeks ( + 2 . 1 ) , and there were 3 (43%) deliveries <37 weeks and no deliveries <34 weeks. Placenta accreta was associated with substantial maternal morbidity. Estimated blood loss exceeded 2000 ml in 41 cases (66%), 5000 ml in 9 (15%), 10,000 ml in 4 (6.5%), and 20,000 ml in 2 (3%). Thirty-two women (55%) required blood transfusions. Blood product replacement exceeded 5 units in 13 (21%), 20 units in 5 (8%), and 70 units in 3 (5%). In three cases ureteral transection and reimplantation or reanastomosis were performed. Cystotomy was performed intentionally in four cases to aid in dissection of the lower uterine segment from the bladder. Additional morbidity included disseminated intravascular coagulation (n = 5), hypotensive shock (n = 2), and enterotomy (n = 1). Two women underwent reoperation for control of hemorrhage. There were no maternal deaths.

Comment
Although earlier reports accepted clinical recognition of abnormal placental adherence as the basis for diagnosis of placenta accreta, ~"4 recent studies have distinguished between suspected and histologically confirmed cases. 1'5-v In this series clinical suspicion of placenta accreta was an unreliable predictor of histologie findings, correctly identifying only 48% of cases. Therefore we confined our analysis to histologically confirmed cases. We recognize that exclusion of unconfirmed cases may underestimate the true incidence of placenta accreta. Moreover, even in the absence of confirmed invasion, low placental implantation may give rise to significant bleeding that requires hysterectomy to avoid maternal death. However, given the uncertainty of the clinical diagnosis, we believe that this approach provides more reliable information and firmly establishes the m i n i m u m incidence of placenta accreta. At our institution the incidence of previous cesarean section has increased from fewer than 8% of deliveries in 1980 to >14% in 1995. The incidence of placenta previa has increased from 0.3% 5 to 0.4% in the current study, and the average age of women with placenta previa has increased from 27 years5 to 30.8 years. Concomitantly with increases in the major clinical risk factors, the

incidence of placenta accreta has risen from 1 in 4027 in 1975 through 19788 to 1 in 2510 in the years encompassed by our report. These observations are consistent with previous reports of a gradual increase in the incidence of placenta accreta. 5' 7, 9 A major contributor to this rise appears to be the increasing incidence of previous cesarean delivery in the wake of increased utilization of the procedure during the last two decades. Cesarean delivery is associated with increased risks of placenta previa5'7 and uterine rupture 1'11 in subsequent pregnancies. Evidence of an association with placenta accreta further emphasizes the potential long-term impact of the decision to perform the initial cesarean section. Eighty percent of confirmed cases of placenta accreta occurred in the 0.4% of the population with placenta previa. The observation that 50% of women with placenta previa had a previous uterine curettage suggests that endometrial disruption or scarring might predispose to abnormal placental implantation. Unfortunately, the incidence of curettage in the remainder of the population is not available to confirm this association. However, among women with placenta previa and a previous cesarean section, the risk of placenta accreta was significantly higher when the placenta was implanted over the scar, supporting the theory that trophoblast adherence or invasion is enhanced when the scant decidualization of the lower uterine segment is further impaired by previous myometrial disruption.4' 5 Contemporary information is limited regarding maternal and perinatal outcome in cases of placenta accreta. In this series, which is the largest report from a single institution, maternal morbidity was substantial and the incidence ofprematurity was high. However, the preterm birth rate with placenta previa-placenta accreta (65%) was similar to that with placenta previa alone (63%), suggesting that the site of placental implantation, rather than the degree of placental invasion, may be the major determinant of the risk of preterm delivery. The incidence of preterm delivery that we observed in women with placenta previa is similar to the rates of 56% to 70% reported by other authors. 12' 1~ Despite increased morbidity, there were no deaths in the present series. T h i s contrasts sharply with the maternal mortality of 9.5% and

AmJ Obstet Gynecol

Volume 177, N u m b e r 1

Miller, Chollet, and Goodwin 213

the fetal death rate of 9.6% reported by Fox 4 in a review of 622 cases of placenta accreta reported between 1945 and 1969. In most cases of placenta accreta the principal concern is control of hemorrhage. In the majority of reported cases this has been achieved by hysterectomy. Fox 4 reported maternal mortality of 5.8% to 6.6% among women who underwent immediate hysterectomy. By comparison, attempts at conservative management were associated with maternal mortality of 12.5% to 28.3%. Although contemporary information regarding conservative management of placenta accreta is limited, in 10 recent reports including 31 cases there were no maternal deaths.5, 8, 14-21 Subsequent fertility did not appear to be impaired. Initial approaches to conservative management of placenta accreta include curettage, oversewing of the placental bed, ligation of the uterine arteries, or ligation of the anterior divisions of the internal iliac arteries. Reported success rates of these maneuvers vary widely. AbdRabbo 22 recently described a stepwise technique of uterine devascularization for management of uncontrollable postpartum hemorrhage. The approach involved sequential ligation of (1) unilateral uterine vessels, (2) bilateral uterine vessels, (3) bilateral low uterine vessels, (4) unilateral ovarian vessels, and (5) bilateral ovarian vessels. Among 103 patients described, the success rate was 100%; however, only two women had placenta previa-placenta accreta. Angiographic embolization of pelvic arteries with gelatin sponge particles or spring coils may provide another option in the management of obstetric hemorrhage. Bakri and Linjawi2~ reported successful control of hemorrhage in all 14 cases of gynecologic or obstetric hemorrhage. Only one of these cases involved placenta accreta. There are several case reports of placenta accreta in which all or part of the placenta was left in situ and managed expectantly with or without methotrexate. 18-21 This approach should be considered only in hemodynamically stable patients who strongly desire preservation of fertility and who understand and accept the risks of delayed hemorrhage and infection. Methotrexate failure la and subsequent successful pregnancy19 have been reported. Conservative options may provide alternatives to hysterectomy in carefully selected patients. However, in the vast majority of cases hysterectomy remains the procedure of choice in the management of placenta accreta. Recently, transvaginal ultrasonography and color Doppler imaging have yielded encouraging results in the prospective diagnosis of placenta accreta. Ultrasonographic findings including loss of the n o r m a l retroplacental hypoechoic zone, thinning or disruption of the

hyperechoic interface between the uterine serosa and the bladder, intraplacental vascular lacunae, and loss of the normal venous flow pattern of the peripheral placental margin have led to accurate diagnoses of placenta accreta in 78% (14/18) to 100% (5/5) of cases. 24' 25 In settings with this diagnostic capability, clinical risk factors may be used as screening criteria to identify women at highest risk for placenta accreta, for whom ultrasonographic evaluation might offer the greatest benefit. In emergency situations or in settings in which ultrasonographic diagnosis is not available, awareness of clinical risk factors can aid in careful preoperative preparation and in counseling women with placenta previa regarding the likelihood of encountering placenta accreta with its attendant morbidity.

REFERENCES

1. Breen JL, Neubecker R, Gregori CA, FranklinJE. Placenta accreta, increta, and percreta. Obstet Gynecol 1977;49:43-7. 2. Williams RW, Creasy RK, Cunningham GC, Hawes WE, Norris FD, Tashiro M. Fetal growth and perinatal viabilityin California. Obstet Gynecol 1982;59:624-32. 3. Kistner RW, Hertig AT, Reid DE. Simultaneously occurring placenta previa and placenta accreta. Surg Gynecol Obstet 1952;94:141-4. 4. Fox H. Placenta accreta, 1945-1969. Obstet Gynecol Surv 1972;27:475-90. 5. Clark SL, Koonings PP, Phelan JP. Placenta previa-accreta and prior cesarean section. Obstet Gynecol 1985;66:89-92. 6. Morison JE. Placenta accreta: a clinicopathologic review of 67 cases. In: Wynn R, editor. Volume 7: obstetrics and gynecology annual. New York: Appleton-Century- Crofts; 1978. p. 107-23. 7. To VVWK, Leung WC. Placenta previa and previous cesarean section. IntJ Gynaecol Obstet 1995;51:25-31. 8. ReadJA, Cotton, DB, Miller FC. Placenta accreta: changing clinical aspects and outcome. Obstet Gynecol 1980;56:31-4. 9. Zelop CM, Harlow BL, Frigoletto FD, Safon LE, Saltzman DH. Emergency peripartum hysterectomy. Am J Obstet Gynecol 1993;168:1443-8. 10. Flamm BL, Newman LA, Thomas SJ, Fallon D, Yoshida MM. Vaginal birth after cesarean delivery: results of a 5-year multicenter collaborative study. Obstet Gynecol 1990;76: 750-4. 11. Miller DA, Diaz FG, Paul RH. Vaginal birth after cesarean: a 10-year experience. Obstet Gynecol 1994;84:255-8. 12. Cotton DB, Read JA, Paul RH, Quilligan EJ. The conservative aggressive management of placenta previa. Am J Obstet Gynecol 1980;137:687-95. 13. McShane PM, Heyl PS, Epstein MF. Maternal and perinatal morbidity resulting from placenta previa. Obstet Gynecol 1985;65:176-82. 14. Komulainen MH, Vayrynen MA, Kauko ML, Saarikoski S. Two cases of placenta accreta managed conservatively.EurJ Obstet Gynecol Reprod Biol 1995;62:135-7. 15. Gorodeski IG, Bahari CM, Holzinger M, Schachter A, Neri A. Placenta previa with focal accretion. Isr J Med Sci 1982;18:277-80. 16. Oumachigui A, Rajagopalan G, Reddy R, Prabhavathy R, Chakravarty A. Placenta accreta and percreta: a review of 5 cases. IntJ Gynaecol Obstet 1981;19:337-40. 17. Hollander DI, Pupkin MJ, Crenshaw MC, Nagey DA. Conservative management of placenta accreta: a case report. J Reprod Med 1988;33:74-8. 18. Jatfe R, DuBeshter B, Sherer DM, Thompson EA, Woods JR

214

Miller, Chollet, and Goodwin

july 1997 Am J Obstet Gynecol

19. 20. 21. 22.

Jr. Failure of methotrexate for term placenta percreta. AmJ Obstet Gynecol 1994;171:558-9. Legro RS, Price FV, Hill LM, Caritis SN. Nonsurgical management of placenta percreta: a case report. Obstet Gynecol 1994;83:847-9. Gibb DMF, Soothill PW, Ward KJ. Conservative management of placenta accreta. Br J Obstet Gynaecol 1994;101: 79-80. Arulkumaran S, Ng CS, Ingemarsson I, Ratnam SS. Medical treatment of placenta accreta with methotrexate. Acta Obstet Gynecol Scand 1986;65:285-6. AbdRabbo SA Stepwise uterine devascularization: a novel technique for management of uncontrollable postpartum

hemorrhage with preservation of the uterus. Am J Obstet Gynecol 1994;171:694-700. 23. Bakri YN, Linjawi T. Angiographic embolization for control of pelvic genital tract hemorrhage: report of 14 cases. Acta Obstet Gynecol Scand 1992;71:1%21. 24. Lerner JP, Dean S, Timor-Tritsch IE. Characterization of placenta accreta using transvaginal sonography and color Doppler imaging. Ultrasound Obstet Gynecol 1995;5:198201. 25. Finberg HJ, Williams JW. Placenta accreta: prospective sonographic diagnosis in patients with placenta previa and prior cesarean section. J Ultrasound Med 1992;11:333-43.

Make researching easier with the Ten-Year Cumtdative Index (1978-1987) to the American Journal of Obstetrics and Gynecology
T h e Ten-Year Cumulative I n d e x is a c o m p l e t e guide to m o r e than 30,000 pages of original articles, case reports, letters, and editorials published in the A m e r i c a n J o u r n a l of Obstetrics and Gynecology f r o m 1978 t h r o u g h 1987 (volumes 130 to 157). N u m b e r i n g m o r e than 800 pages, the Cumulative I n d e x includes b o t h a subject and an a u t h o r index. This hard-cover v o l u m e is $64.50 ($69.00 international). Prices include shipping. Payment must accompany all
orders and must be in U.S. funds, drawn on a U.S. bank.

Contact Mosby-Year Book, Inc., Subscription Services, 11830 Westline Industrial Drive, St. Louis, MO 63146-3318, USA; p h o n e (314)453-4351. In the U.S., call toll-free: (800)453-4351. Please allow six weeks for delivery.