Hemorrhage is second only to embolism as a cause of pregnancy-related mortality in the United States.1 Much of the existing data on cesarean-associated transfusion was retrospectively collected and is not contemporary. In this investigation, we have more precisely defined risk factors for the administration of packed red blood cells in association with cesarean delivery in a recently assembled, large prospective cohort. Our hope was that these data, reflective of current transfusion practices in a broadly representative sample of U.S. women, might be useful in patient counseling, perioperative planning, and even perhaps in reducing the risk of transfusion.
MATERIALS AND METHODS
Subjects were identified as part of a 19-center observational study of cesarean delivery for the four-year period January 1, 1999, to December 31, 2002. Each clinical center and the data coordinating center had institutional review board approval for this study, and the requirement for informed consent was waived. For the first 2 years of the study, women who underwent cesarean delivery were identified on a daily basis by trained, certified research personnel who abstracted data from hospital records and logs onto standardized forms. For the final 2 years of the study, only women who underwent repeat cesarean delivery were identified and had their data collected. This change occurred because the primary focus of this observational study was trial of labor after cesarean. Although planned as only a 2-year study, the declining rate of trial of labor after cesarean necessitated 2 more years of data collection to meet the sample size goal for this group. Gestational age was assigned based on the best estimate of the local caregivers (and in 28%, 46%, and 11%, respectively, included a 1st, 2nd, or 3rd trimester ultrasound). In all of the participating centers, some form of prophylaxis against uterine atony was routinely employed. However, we did not collect information on specific prophylactic regimens for individual women.
Data lacking unique patient identifiers were transmitted weekly from each of the 19 clinical centers by telecommunications link to the data coordinating center where they were edited for missing, out of range, and inconsistent values. Weekly, edit reports were transmitted to each center for correction or clarification. Data were also compared across forms at regular intervals, and corrections and clarifications were requested from the centers as appropriate.
The independent variable for this study was transfusion of packed red blood cells, intraoperatively or postoperatively before hospital discharge (“transfusion”). Because the 1) obstetric characteristics and clinical situations in which primary and repeat cesareans were performed differed substantially (eg, the former most often experienced labor and the latter did not); 2) outcomes associated with certain conditions (eg, placenta previa) were predicted to be dependent on whether the cesarean was a primary or a repeat; and 3) data collection period for the two conditions was not synchronous, we analyzed primary and repeat cesarean deliveries separately.
Categorical variables were compared by the χ2 test, Fisher exact test, or Mantel-Haenszel test of trend, as appropriate. Continuous variables were compared using the Wilcoxon rank sum test. Variables found to be significant in univariable analysis were entered into a multivariable logistic regression model. Backward selection was used to retain only those variables with P values less than .05. Nominal two-tailed P values are reported with no adjustments made for multiple comparisons. For this analysis a P value less than .05 was considered statistically significant. All analyses of multiple gestation were predicated only on the first infant delivered. SAS 8 software (SAS Institute, Inc., Cary, NC) was used for analysis.
A total of 23,486 pregnancies were analyzed. Seven hundred sixty-two women (3.2%) were transfused (median 2 units, 25th to 75th % 2–3 units). In 114 (15%) women, the transfusion was given only intraoperatively, in 556 (73%) it was given only postoperatively, and in 92 (12%) women it was given during both periods. The mean (plus or minus standard deviation) maternal age was 27 (±7) years. Forty-two percent were white, 31% African American, 21% Hispanic, and 5% other. Most (68%) were nulliparous. Mean body mass index (BMI) at delivery was 32.5 kg/m2 (±7.1). Fifty-five percent had government-funded (or no) insurance and 45% private insurance. Mean gestational age at delivery was 38 (±4) weeks, and mean birth weight 3,039 (±943) gm. Seventy-five percent experienced labor. The predominant cesarean indication was abnormally progressive labor (37%), followed by nonreassuring fetal heart rate status (23%).
In univariable analyses, several continuous variables were significantly associated with transfusion (Table 1). Compared with women who did not undergo transfusion, women who did were, on average, approximately one-half year younger, had a slightly lower BMI, and had a preoperative hematocrit that was 4% lower. Their mean gestational age at delivery was nearly 2 weeks earlier; if induced, their median induction time was 2 hours longer; and if they received oxytocin, it was administered for almost 2 hours longer. The average birth weight of their infants was 338 g lighter. We also analyzed hematocrit as a continuous variable: for each percentage point increase in hematocrit, the odds ratio (OR) for transfusion was 0.8 (95% confidence interval [CI] 0.79–0.82).
In univariable analyses of categorical variables, significant risk factors for transfusion included nonwhite race, multiparity (and progressively higher parity), multiple gestation, hypertensive disorders, clinically diagnosed chorioamnionitis,2 placental abruption, placenta previa, general anesthesia, mild (preoperative hematocrit 25–29%) and severe (preoperative hematocrit less than 25%) anemia, hysterotomy other than low transverse, and prematurity (gestational age less than 37 weeks, Table 2).
We combined factors from the above univariable analyses into a multivariable logistic regression model. In the final multivariable model, African American or Hispanic race, multiple gestation, preeclampsia, chorioamnionitis, placental abruption, mild preoperative maternal anemia (hematocrit 25–29%), general anesthesia, eclampsia or hemolysis, elevated liver enzymes, low platelets syndrome, placenta previa, and severe preoperative maternal anemia (hematocrit less than 25%) were significantly associated with transfusion in a progressively stronger fashion (Table 3).
A total of 33,683 pregnancies were analyzed. Seven hundred thirty-five women (2.2%) were transfused (median 2 units, 25th% to 75th% 2–4 units). In 169 (23%) women, the transfusion was given only intraoperatively, in 436 (59%) it was given only postoperatively, and in 130 (18%) women it was given during both periods. The mean (plus or minus standard deviation) maternal age was 30 (±6) years. Thirty-nine percent were white, 26% African American, 31% Hispanic, and 4% other. Median parity was 1. Mean BMI at delivery was 33.4 kg/m2 (±7.2). Fifty-seven percent had government-funded (or no) insurance, and 43% had private insurance. Mean gestational age at delivery was 38 (±3) weeks, and mean birth weight 3261 (±723) gm. Twenty-nine percent experienced labor. The predominant cesarean indication was elective repeat (63%).
In univariable analyses, several continuous variables were significantly associated with transfusion (Table 4). Compared with women who did not undergo transfusion, women who underwent transfusion were, on average, of lower BMI, and had a preoperative hematocrit that was almost 4% lower. Their mean gestational age at delivery was almost 2 weeks earlier and the mean birth weight of their infants was 427 g lighter. If their labor was augmented, the duration of augmentation was shorter. For each percentage point increase in hematocrit, the OR for transfusion was 0.8 (95% CI 0.77–0.80).
In univariable analyses of categorical variables, significant risk factors for transfusion included nonwhite race, parity of 3 and above, multiple gestation, three or more prior cesareans, hypertensive disorders, clinically diagnosed chorioamnionitis, placental abruption, placenta previa, general anesthesia, mild (preoperative hematocrit 25–29%) and severe (preoperative hematocrit less than 25%) anemia, hysterotomy other than low transverse, and prematurity (Table 5).
We combined significant factors from the above univariable analyses into a multivariable regression model and found preeclampsia, African-American or Hispanic race, race category “other,” chorioamnionitis, placental abruption, mild preoperative maternal anemia (hematocrit 25–29%), general anesthesia, 5 or more prior cesareans, placenta previa, and severe preoperative maternal anemia (hematocrit less than 25%) were significantly associated with transfusion in a progressively stronger fashion (Table 6).
In this contemporary, prospectively assembled cohort of approximately 57,000 women undergoing cesarean delivery, we have confirmed that overall the risk of cesarean-associated packed red blood cell transfusion is relatively low: 3.2% for primary cesarean and 2.2% for repeat. These rates are comparable to the contemporaneous rate of 3.3% (primary and repeat cesareans combined) reported in a four-hospital Danish study and a single hospital study in which a rate of 3.2% was reported for 1987.3,4
In addition to the large size of the cohort, a major strength of this study is that all data for the project were abstracted by trained and certified research nurses before hospital discharge using uniform criteria, and ongoing, systematic procedures for data quality control were employed. The major weakness of these data is that they are observational and do not reflect standardized transfusion practices. Nor did the analyses we performed take into account specific prophylactic regimens for the prevention of uterine atony, because this information was not collected.
Our analysis has revealed multiple factors that are associated with an increased risk of transfusion for women undergoing both a primary and a repeat cesarean delivery. Some of these factors have been previously identified, although from much smaller cohorts.5 Most notably, among women who underwent primary cesarean, a preoperative hematocrit of less than 25% was associated with a 36% risk of transfusion. Among women who underwent repeat cesarean, this degree of anemia also was associated with a high risk of transfusion (28%), as was the occurrence of placenta previa (32% risk). These risks argue for optimizing maternal antenatal iron status to avoid severe anemia, because iron deficiency is the most common cause of anemia during pregnancy,6 and they suggest that informing severely anemic iron deficient women about their high risk of transfusion should they undergo cesarean might enhance compliance with iron supplementation. These data also argue for careful perioperative planning when placenta previa complicates repeat cesarean.
As have others, we found that nontransverse hysterotomy was associated with an increased risk of transfusion, but with multivariable analysis this association was not statistically significant.7 Also as have others, we found that general anesthesia was independently associated with an increased risk of transfusion.8–10 In our study, the odds were quadrupled for women undergoing primary cesarean, and increased seven-fold for women undergoing repeat cesarean. In addition to the well-recognized risk of aspiration and failed intubation associated with general anesthesia, our data suggest that, all other things being equal, avoidance of transfusion, especially in the face of severe anemia, may be another reason to choose regional anesthesia for cesarean.
1. Chang J, Elam-Evans LD, Berg CJ, Herndon J, Flowers L, Seed KA, et al. Pregnancy-related mortality surveillance—United States, 1991-1999. MMWR Surveill Summ 2003;52:1–8.
2. Rouse DJ, Landon M, Leveno KJ, Leindecker S, Varner MW, Caritis SN, et al. The Maternal-Fetal Medicine Units cesarean registry: chorioamnionitis at term and its duration—relationship to outcomes. Am J Obstet Gynecol 2004;191:211–6.
3. Larsen R, Titlestad K, Lillevang ST, Thomsen SG, Kidholm K, Georgsen J. Cesarean section: is pretransfusion testing for red cell alloantibodies necessary? Acta Obstet Gynecol Scand 2005;84:448–55.
4. Camann WR, Datta S. Red cell use during cesarean delivery. Transfusion 1991;31:12–5.
5. Cousins LM, Teplick FB, Poeltler DM. Pre-cesarean blood bank orders: a safe and less expensive approach. Obstet Gynecol 1996;87:912–6.
6. Cunningham FG, Leveno KJ, Bloom SL, Hauth JC, Gilstrap LC III, Wenstrom KD, editors Williams obstetrics. 22nd ed. New York (NY): McGraw-Hill Professional; 2005. p. 1145.
7. Patterson LS, O'Connell CM, Baskett TF. Maternal and perinatal morbidity associated with classic and inverted T cesarean incisions. Obstet Gynecol 2002;100:633–7.
8. Combs CA, Murphy EL, Laros RK Jr. Factors associated with hemorrhage in cesarean deliveries. Obstet Gynecol 1991;77: 77–82.
9. Gilstrap LC 3rd, Hauth JC, Hankins GD, Patterson AR. Effect of type of anesthesia on blood loss at cesarean section. Obstet Gynecol 1987;69:328–32.
10. Hager RM, Daltveit AK, Hofoss D, Nilsen ST, Kolaas T, Oian P, et al. Complications of cesarean deliveries: rates and risk factors. Am J Obstet Gynecol 2004;190:428–34.
In addition to the authors, other members of the National Institute of Child Health and Human Development Maternal-Fetal Medicine Units Network are as follows:
Ohio State University—J. Iams, F. Johnson, S. Meadows, H. Walker
University of Alabama at Birmingham—J. Hauth, A. Northen, S. Tate
University of Texas Southwestern Medical Center—S. Bloom, J. McCampbell, D. Bradford
University of Utah—M. Belfort, F. Porter, B. Oshiro, K. Anderson, A. Guzman
University of Chicago—J. Hibbard, P. Jones, M. Ramos-Brinson, M. Moran, D. Scott
University of Pittsburgh—K. Lain, M. Cotroneo, D. Fischer, M. Luce
Wake Forest University—M. Harper, M. Swain, C. Moorefield, K. Lanier, L. Steele
Thomas Jefferson University—A. Sciscione, M. DiVito, M. Talucci, M. Pollock
Wayne State University—M. Dombrowski, G. Norman, A. Millinder, C. Sudz, B. Steffy
University of Cincinnati—T. Siddiqi, H. How, N. Elder
Columbia University—F. Malone, M. D'Alton, V. Pemberton, V. Carmona, H. Husami
Brown University—H. Silver, J. Tillinghast, D. Catlow, D. Allard
Northwestern University—M. Socol, D. Gradishar, G. Mallett
University of Miami— G. Burkett, J. Gilles, J. Potter, F. Doyle, S. Chandler
University of Tennessee—W. Mabie, R. Ramsey
University of Texas at San Antonio—O. Langer, S. Barker, M. Rodriguez
University of North Carolina—K. Moise, K. Dorman, S. Brody, J. Mitchell
University of Texas at Houston—L. Gilstrap, M. Day, M. Kerr, E. Gildersleeve
Case Western Reserve University—H. Ehrenberg, C. Milluzzi, B. Slivers, C. Santori
The George Washington University Biostatistics Center—E. Thom, S. Gilbert, H. Juliussen-Stevenson, M. Fischer
National Institute of Child Health and Human Development—D. McNellis, K. Howell, S. Pagliaro