Original Research

Blood Transfusion During Pregnancy, Birth, and the Postnatal Period Jillian A. Patterson, BScAdv(Hons), MBiostat, Christine L. Roberts, MBBS, DrPH, Jennifer R. Bowen, MBBS (Hons), FRACP, David O. Irving, MSc, PhD, James P. Isbister, BSc(Med) Hons, MB BS Hons, Jonathan M. Morris, MBChB, PhD, and Jane B. Ford, BA(Hons), PhD OBJECTIVES: To identify risk factors for transfusion and trends in transfusion rates across pregnancy and the postnatal period. METHODS: Linked hospital and birth data on all births in hospitals in New South Wales, Australia, between 2001 and 2010 were used to identify blood transfusions for women during pregnancy, at birth, and in the 6 weeks postpartum. Poisson regression was used to identify risk factors for red cell transfusion in the birth admission. Separate models were fitted for cesarean and vaginal births. RESULTS: Between 2001 and 2010, there were 12,147 transfusions across 891,914 pregnancies, with a transfusion rate of 1.4%. The transfusion rate increased steadily from 1.2% in 2001 to 1.6% in 2010. The majority of transfusions (91%) occurred during the birth admission, and 81% of these transfusions were associated with a diagnosis of hemorrhage. Women with bleeding or From the Kolling Institute, University of Sydney, the Royal North Shore Hospital, the University of Sydney, and the Australian Red Cross Blood Service, Sydney, New South Wales, Australia. Supported by a Partnership Grant from the Australian National Health and Medical Research Council NHMRC (#1027262), the Australian Red Cross, and the NSW Clinical Excellence Commission. Christine L. Roberts is supported by a NHMRC Senior Research Fellowship (#1021025). Jane B. Ford is supported by an ARC Future Fellowship (#120100069). The authors thank the NSW Ministry of Health for access to the population health data and the NSW Centre for Health Record Linkage for linking the data sets. An earlier version of these findings was presented in poster form at the Society for Pediatric and Perinatal Epidemiologic Research 26th Annual Meeting, June 17– 18, 2013, Boston, Massachusetts. Corresponding author: Jillian A. Patterson, BScAdv(Hons), MBiostat, Clinical & Population Perinatal Health Research, c/o University Department of O&G, Building 52, Royal North Shore Hospital, St Leonards, New South Wales 2065, Australia; e-mail: [email protected]. Financial Disclosure The authors did not report any potential conflicts of interest. © 2013 by The American College of Obstetricians and Gynecologists. Published by Lippincott Williams & Wilkins. ISSN: 0029-7844/14

126

VOL. 123, NO. 1, JANUARY 2014

platelet disorders (vaginal: number transfused 529, relative risk [RR] 7.8, 99% confidence interval [CI] 6.9–8.7, cesarean: n5592, RR 8.7, CI 7.7–9.7) and placenta previa: (vaginal n573, RR 4.6, CI 3.4–6.3, cesarean: n5875, RR 5.7, CI 5.1–6.4) were at highest risk of transfusion. Among vaginal births, increased risk was evident for forceps (n51,036, RR 2.8, CI 2.5–3.0) or vacuum births (n51,073, RR 1.9, CI 1.7–2.0) compared with nonoperative births. CONCLUSIONS: Rates of obstetric blood product transfusion have increased by 33% since 2001, with the majority of this associated with hemorrhage. Women with bleeding or platelet disorders and placenta previa are at increased risk of transfusion and should be treated accordingly. (Obstet Gynecol 2014;123:126–33) DOI: 10.1097/AOG.0000000000000054

LEVEL OF EVIDENCE: II

B

lood for transfusion is a limited, costly resource, and its use has specific risks.1,2 Internationally, this has led to increased efforts to reduce unnecessary blood use across many disciplines.3 Transfusion of red blood cells generally has decreased in Australia in recent years4; however, there is evidence of increasing rates of maternal red blood cell transfusion around childbirth.5 This trend has also been observed in the United States, Canada, Finland, and Ireland, particularly in the context of postpartum hemorrhage.6–12 Between 1998 and 2009 in the United States, there was a steep increase in transfusions during a delivery admission (from 0.3% to 1.0%).12 The obstetric blood transfusion rate in Australia in 2002 was 0.88%, which was higher than contemporaneous rates reported in other countries, including the United States (0.46% in 2003),7 Canada (0.63% in 2004),13 and Ireland (0.84% in 2003).11 Reasons for the higher rate are unknown, although they may relate to differences in

OBSTETRICS & GYNECOLOGY

MATERIALS AND METHODS The study population included women giving birth in New South Wales hospitals between January 2001 and December 2010. New South Wales is the most populous state in Australia, with more than 7 million residents and approximately 90,000 births per year. “Birth data” including maternal characteristics, pregnancy, and birth were obtained from the Perinatal Data Collection. The Perinatal Data Collection is a statutory population-based collection of all births in New South Wales of at least 20 weeks of gestation or 400 g birth weight. These data were linked to “hospital data” from the Admitted Patients Data Collection. The Admitted Patients Data Collection is a census of all hospital discharges in New South Wales and records information on diagnoses and procedures associated with these discharges. Up to 20 of the 55 available diagnoses and procedures for each discharge are coded according to the International Classification of Diseases, 10th Revision, Australian Modification and the Australian Classification of Health Interventions. The New South Wales Centre for Health Record Linkage performed probabilistic data linkage between the two data sets.18 For this study, rates of incorrect and missed links were less than 5 per 1,000. Hospital admissions were classified as antenatal, birth, or postnatal to allow for examination of the different blood product use at each stage of pregnancy. Antenatal admissions were those that occurred

VOL. 123, NO. 1, JANUARY 2014

from 20 weeks of gestation and ended with the mother being discharged before the birth. Birth admissions were separated into those with a prolonged antenatal component (greater than 4 days) and those admitted 4 days or less before delivery. Postnatal admissions were admissions occurring after the initial discharge of the mother but within 6 weeks of the birth. Admissions involving transfusion of blood products were identified from the hospital data using Australian Classification of Health Interventions procedure codes.19 “Blood transfusion” refers to the administration of packed red cells or whole blood and “platelets and coagulation factors” include platelets, coagulation factors, and other serum (including fresh-frozen plasma); “blood products” is used to refer to transfusion of either or both of these. Blood transfusion (sensitivity 83.1%, specificity 99.9%) and administration of platelets and coagulation factors (sensitivity 73.1%, specificity 100%) are well ascertained in the hospital data.17 The quality of reporting of use of other bloodderived products (including leukocytes, gamma globulin) is unknown and so use of these products was not considered. Women with iron deficiency anemia were identified from the hospital data and women with bleeding or platelet disorders were identified by International Classification of Diseases, 10th Revision, Australian Modification codes for conditions such as thalassemia, hemolytic and aplastic anemias, coagulation defects, and idiopathic thrombocytopenic purpura (see the Appendix, available online at http://links.lww.com/AOG/A454). Antepartum hemorrhage included placental abruption or other antepartum hemorrhage. Women without Packed cells only Packed cells and platelets or coagulation factors Platelets or coagulation factors only Transfusion rate per 1,000 deliveries

data collection. Obstetric transfusions tend to be urgent, unpredictable, and occur in otherwise healthy women.14 A small number of population-based studies have identified risk factors for blood transfusion in the maternity setting, including mode of delivery, placenta previa, antepartum hemorrhage, anemia, multiple pregnancies, and the extremes of maternal age.10,15,16 Population data provide a valuable source of data for studies of trends and risk factors for uncommon outcomes such as transfusion.17 These studies typically look only at the birth admission and are unable to estimate the transfusion burden associated with antenatal and postpartum hospitalizations. Few studies have considered blood products other than red cells. This study aims to explore recent trends in blood and blood product transfusion and the use of blood products throughout pregnancy and the postnatal period in women whose pregnancy ended in a registered birth (beyond 20 weeks of gestation). We also examine the risk factors for transfusion during the birth admission.

16 14 12 10 8 6 4 2 0 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Year

Fig. 1. Rate of transfusions per 1,000 deliveries, New South Wales 2001–2010. The transfusion rate increased significantly over this time period overall (P,.001), for packed cells alone (P5.003), and packed cells in conjunction with other products (P,.001) but not for other blood products alone (P5.1). This includes any transfusion during the antenatal period, birth admission, and postnatally. Patterson. Obstetric Blood Transfusion. Obstet Gynecol 2014.

Patterson et al

Obstetric Blood Transfusion

127

Table 1. Characteristics of Pregnancies by Blood Product Transfusion Status During Pregnancy, New South Wales, 2001–2010 Variable Total pregnancies Maternal age (y) Younger than 20 20–34 35 or older Private patient Yes No Smoker Yes No Multiple birth Yes No Nulliparous Yes No Previous cesarean delivery Yes No Bleeding or platelet disorders Yes No Antepartum hemorrhage Yes No Placenta previa Yes No Iron deficiency anemia Yes No Gestational age at birth (wk) 20–32 33–36 37 or more Delivery type Normal vaginal Cesarean Prelabor Intrapartum Instrumental Forceps Vacuum Induction Yes No Birth weight (for gestation) Small

Packed Red Cells or Blood

Platelets or Coagulation Factors

11,760 (100.0)

1,648 (100.0)

612 (5.2) 8,392 (71.4) 2,756 (23.4)

42 (2.5) 1,084 (65.8) 522 (31.7)

626 (5.2) 8,682 (71.5) 2,839 (23.4)

33,500 (3.8) 658,286 (74.8) 187,981 (21.4)

18 (16.7–20.0) 13 (12.7–13.3) 15 (14.2–15.5)

2,602 (22.1) 9,158 (77.9)

472 (28.6) 1,176 (71.4)

2,724 (22.4) 9,423 (77.6)

296,410 (33.7) 583,357 (66.3)

9 (8.7–9.5) 16 (15.5–16.3)

1,963 (16.7) 9,767 (83.1)

197 (12.0) 1,442 (87.5)

2,002 (16.5) 10,113 (83.3)

121,511 (13.8) 756,086 (85.9)

16 (15.4–17.0) 13 (12.9–13.5)

612 (5.2) 11,148 (94.8)

92 (5.6) 1,556 (94.4)

631 (5.2) 11,516 (94.8)

13,383 (1.5) 866,384 (98.5)

45 (41.0–49.1) 13 (12.8–13.4)

5,608 (47.7) 6,136 (52.2)

738 (44.8) 907 (55.0)

5,812 (47.8) 6,317 (52.0)

369,525 (42.0) 509,032 (57.9)

15 (15.0–16.0) 12 (11.9–12.6)

1,726 (14.7) 10,034 (85.3)

361 (21.9) 1,287 (78.1)

1,792 (14.8) 10,355 (85.2)

119,573 (13.6) 760,194 (86.4)

15 (14.0–15.6) 13 (13.1–13.7)

1,132 (9.6) 10,628 (90.4)

291 (17.7) 1,357 (82.3)

1,275 (10.5) 10,872 (89.5)

7,713 (0.9) 872,054 (99.1)

142 (133.3–150.4) 12 (12.0–12.6)

1,045 (8.9) 10,715 (91.1)

238 (14.4) 1,410 (85.6)

1,084 (8.9) 11,063 (91.1)

22,179 (2.5) 857,588 (97.5)

47 (43.4–49.8) 13 (12.5–13.0)

962 (8.2) 10,798 (91.8)

228 (13.8) 1,420 (86.2)

979 (8.1) 11,168 (91.9)

8,485 (1.0) 871,282 (99.0)

103 (96.2–110.7) 13 (12.4–12.9)

715 (6.1) 11,045 (93.9)

54 (3.3) 1,594 (96.7)

721 (5.9) 11,426 (94.1)

4,654 (0.5) 875,113 (99.5)

134 (123.3–145.0) 13 (12.6–13.2)

736 (6.3) 1,182 (10.1) 9,842 (83.7)

207 (12.6) 283 (17.2) 1,158 (70.3)

791 (6.5) 1,266 (10.4) 10,090 (83.1)

13,436 (1.5) 42,836 (4.9) 823,495 (93.6)

56 (51.1–60.1) 29 (26.9–30.6) 12 (11.8–12.4)

537,476 246,130 141,349 104,779 96,308 32,663 63,645

10 18 17 20 22 32 17

5,226 4,343 2,319 2,023 2,176 1,068 1,108

(44.4) (36.9) (19.7) (17.2) (18.5) (9.1) (9.4)

3,794 (32.3) 7,966 (67.7)

895 (7.6)

491 947 583 364 206 102 104

(29.8) (57.5) (35.4) (22.1) (12.5) (6.2) (6.3)

540 (32.8) 1,108 (67.2)

165 (10.0)

Any Blood Products

No Blood Products

12,147 (100.0) 879,767 (100.0)

5,341 4,588 2,484 2,103 2,202 1,081 1,121

(44.0) (37.8) (20.4) (17.3) (18.1) (8.9) (9.2)

(61.1) (28.0) (16.1) (11.9) (10.9) (3.7) (7.2)

Rate/1,000 Deliveries (99% CI) 14 (13.3–13.9)

(9.5–10.2) (17.7–18.9) (16.5–18.1) (18.7–20.7) (21.3–23.4) (29.8–34.3) (16.1–18.5)

3,910 (32.2) 8,237 (67.8)

236,932 (26.9) 642,835 (73.1)

16 (15.6–16.8) 13 (12.3–13.0)

959 (7.9)

83,218 (9.5)

11 (10.5–12.2) (continued )

128

Patterson et al

Obstetric Blood Transfusion

OBSTETRICS & GYNECOLOGY

Table 1. Characteristics of Pregnancies by Blood Product Transfusion Status During Pregnancy, New South Wales, 2001–2010 (continued ) Variable

Packed Red Cells or Blood

Platelets or Coagulation Factors

Any Blood Products

No Blood Products

Rate/1,000 Deliveries (99% CI)

9,039 (76.9) 1,826 (15.5)

1,294 (78.5) 189 (11.5)

9,333 (76.8) 1,855 (15.3)

703,661 (80.0) 92,888 (10.6)

13 (12.8–13.4) 20 (18.5–20.6)

6,157 2,906 1,522 1,562

358,695 188,394 119,278 213,400

17 15 13 7

Average Large Hospital of birth Tertiary Regional Urban or other Private

5,934 2,844 1,488 1,494

(50.5) (24.2) (12.7) (12.7)

982 210 182 274

(59.6) (12.7) (11.0) (16.6)

(50.7) (23.9) (12.5) (12.9)

(40.8) (21.4) (13.6) (24.3)

(16.4–17.4) (14.5–15.8) (11.9–13.3) (6.8–7.7)

CI, confidence interval. Denominator5pregnancies.

bleeding disorders, antepartum hemorrhage, placenta previa, hypertension, and diabetes were considered to have no prior indication for transfusions. Age and gestational age groups were determined based on clinical relevance and women were classified as private patients if they received private obstetric care in a public or private hospital. Rates were calculated per 1,000 deliveries and proportions are proportion of deliveries unless otherwise specified. Multiple births (twins or higher-order multiples) were counted as a single delivery. Trends were assessed using the Cochran Armitage test for trend. Significance was set at a50.01. A multivariable Poisson regression model with robust variances was used to identify factors associated with higher use of blood product transfusions; factors that were significant with a 0.2 in a univariate model were included in the initial multivariable model and removed in a stepwise fashion until only variables significant at a 0.01 remained. This was performed separately for vaginal and cesarean deliveries to account for possible differences in decision-making based on physical location (operating room compared with birth unit) and differing criteria for postpartum hemorrhage (a common indication for blood transfusion). In the International Classification of Diseases, 10th Revision, Australian Modification, postpartum hemorrhage is defined as blood loss of 750 mL or more after cesarean delivery or 500 mL or more after vaginal birth.20 Women with missing data for possible confounding factors were excluded from this analysis. All analyses were performed in SAS 9.3. Ethical approval was obtained from the New South Wales Population and Health Services Research Ethics Committee.

RESULTS In the period 2001–2010, there were 891,914 deliveries to 578,207 women involving 1,117,939 admis-

VOL. 123, NO. 1, JANUARY 2014

sions. The blood product transfusion rate was 1.4% (99% confidence interval [CI] 1.3–1.4) of deliveries with 11,529 mothers receiving a transfusion in 12,147 pregnancies or the postnatal period. During the time period, 286 women had more than one delivery involving a transfusion, including 50 (17%) women with a bleeding disorder. Blood products were transfused in 484 antenatal admissions, 667 prolonged birth admissions, 10,715 birth admissions, and 600 postnatal admissions. The transfusion rate was highest for the birth admission (12.0/1,000 deliveries) compared with the antenatal (0.5/1,000 deliveries), prolonged birth admissions (0.87/1,000 deliveries), and postnatal admissions (0.7/1,000 deliveries). Ninetyone percent (11,382) of transfusions occurred in the birth admission, whereas 3.9% were antenatal and 4.8% were postnatal transfusions. The transfusion of blood products at any stage during pregnancy, birth, or the postnatal period increased steadily from 1.2% in 2001 to 1.6% in 2010 (P#.001) (Fig. 1). When considering only birth admission, the rate increased from 11.0 per 1,000 in 2001 to 15.3 per 1,000 in 2010 (P,.001). There has been little change in the type of products used with the majority of women (86%) receiving only packed red cells, whole blood, or both, packed cells making up the majority of this (99.4%). There was a significant increase in the number of women receiving packed cells alone (P5.003) or in conjunction with other blood products (P,.001), but not in those receiving other products alone (P5.1). When compared with red cell use, platelets and coagulation factors were more commonly used among women aged 35 years and older, private patients, multiparous women, and those having a cesarean delivery (Table 1). Where blood product transfusion occurred during pregnancy, this usually occurred in a single admission (98.6%) with less than 0.4% of deliveries involving three or more admissions involving transfusions.

Patterson et al

Obstetric Blood Transfusion

129

Table 2. Factors Associated With Blood or Blood Product Transfusion in the Birth Admission, New South Wales, 2001–2010 Vaginal Deliveries

Factor Maternal age (y) Younger than 20 20–35 35 or older Private patient Smoker Australian-born Multiple birth Nulliparous Previous cesarean delivery Bleeding or platelet disorder Iron deficiency anemia Chronic hypertension Pregnancy hypertension Any diabetes Placenta previa Antepartum hemorrhage Augmentation of labor Induction Intrapartum cesarean delivery Prolonged labor Regional analgesia Birth weight (for gestation) Small Average Large Gestational age (wk) Less than 33 33–36 37 or more Delivery type Vaginal Forceps Vacuum Year of birth Hospital type Tertiary Metropolitan Regional Private

Transfused

Not Transfused

n (%)

n (%)

427 5,210 1,268 1,386 1,105 4,587 230 3,719 377

(6.2) (75.5) (18.4) (20.1) (16.0) (66.4) (3.3) (53.9) (5.5)

28,076 484,563 116,070 187,206 93,757 444,368 4,857 257,193 22,879

Cesarean Deliveries

Multivariable Model Relative Risk

99% CI

1.2 1 1.1 0.8 1.1 0.8 2.8 1.4 1.8

(1.02–1.34) (Reference) (1.00–1.18) (0.68–0.86) (1.04–1.25) (0.75–0.87) (2.36–3.41) (1.32–1.54) (1.58–2.09)

(4.5) (77.1) (18.5) (29.8) (14.9) (70.7) (0.8) (40.9) (3.6)

Transfused

Not Transfused

n (%)

n (%)

154 2,820 1,366 1,146 694 2,940 356 1,746 1,297

(3.5) (65.0) (31.5) (26.4) (16.0) (67.7) (8.2) (40.2) (29.9)

5,239 169,740 70,456 107,489 27,103 174,578 8,314 110,360 96,415

(2.1) (69.2) (28.7) (43.8) (11.0) (71.1) (3.4) (45.0) (39.3)

Multivariable Model Relative Risk

99% CI

1.5 1 1.2 0.8 — 0.9 1.7 0.7 0.7

(1.23–1.87) (Reference) (1.07–1.27) (0.67–0.85) — (0.80–0.96) (1.44–1.98) (0.60–0.74) (0.66–0.81)

529 (7.7)

4,967 (0.8)

7.8

(6.93–8.73)

592 (13.6)

2,833 (1.2)

8.7

(7.69–9.76)

344 (5.0)

3,337 (0.5)

—

—

258 (5.9)

1,399 (0.6)

—

—

75 (1.1)

4,235 (0.7)

—

—

99 (2.3)

3,687 (1.5)

—

—

1,009 (14.6)

48,872 (7.8)

1.6

(1.45–1.74)

884 (20.4)

32,244 (13.1)

1.5

(1.38–1.68)

434 (6.3) 73 (1.1) 354 (5.1)

32,426 (5.2) 939 (0.1) 13,157 (2.1)

— 4.6 1.9

— (3.44–6.26) (1.63–2.16)

388 (8.9) 875 (20.2) 656 (15.1)

20,623 (8.4) 7,495 (3.1) 8,662 (3.5)

— 5.7 2.2

— (5.12–6.40) (1.96–2.51)

988 (14.3)

63,152 (10.0)

1.3

(1.17–1.42)

260 (6.0)

15,519 (6.3)

—

—

2,756 (39.9) 188,769 (30.0) * *

1.5 *

(1.36–1.56) 860 (19.8) 46,537 (19.0) * 1,996 (46.0) 104,563 (42.6)

1.1 1.3

(1.01–1.28) (1.21–1.47)

665 (9.6) 28,522 (4.5) 2,194 (31.8) 149,589 (23.8)

— —

— *

— *

508 (7.4) 60,251 (9.6) 5,389 (78.0) 510,541 (81.2) 1,008 (14.6) 57,917 (9.2)

0.7 1 1.7

(0.65–0.82) 357 (8.2) 22,286 (9.1) (Reference) 3,246 (74.8) 188,551 (76.8) (1.57–1.88) 737 (17.0) 34,598 (14.1)

0.8 1 1.3

(0.71–0.94) (Reference) (1.15–1.42)

186 (2.7) 6,218 (1.0) 403 (5.8) 24,933 (4.0) 6,316 (91.5) 597,558 (95.0)

1.8 1.2 1

(1.49–2.21) 488 (11.2) 5,440 (2.2) (1.02–1.35) 774 (17.8) 17,425 (7.1) (Reference) 3,078 (70.9) 222,570 (90.7)

2.7 1.7 1

(2.36–3.15) (1.53–1.93) (Reference)

4,850 (70.2) 536,014 (85.3) 1,036 (15.0) 32,338 (5.1) 1,073 (15.5) 63,364 (10.1)

1 2.8 1.6 1.0

(Reference) (2.51–3.04) (1.50–1.81) (1.04–1.06)

* * * 1.0

* * * (1.02–1.05)

3,409 1,718 971 807

1.1 1 1.2 0.6

(0.98–1.18) 2,312 (53.3) (Reference) 971 (22.4) (1.11–1.38) 425 (9.8) (0.53–0.73) 632 (14.6)

1.1 1 1.3 0.6

(0.93–1.22) (Reference) (1.09–1.48) (0.53–0.78)

†

†

(49.4) 260,617 (41.5) (24.9) 141,423 (22.5) (14.1) 94,078 (15.0) (11.7) 132,591 (21.1)

— —

145 (3.3) *

* * * †

5,102 (2.1) *

* * * †

94,631 46,050 24,725 80,029

(38.6) (18.8) (10.1) (32.6)

CI, confidence interval. Reference categories are absence of risk factor unless otherwise specified. — Indicates that risk factor is not retained in multivariable model. * Adjustment for this risk factor was not applicable for this mode of delivery. † Year treated as a continuous variable; for trends, see Figure 1.

130

Patterson et al

Obstetric Blood Transfusion

OBSTETRICS & GYNECOLOGY

Rates of transfusion in the birth and postnatal period were high in women having a hysterectomy in the birth admission (896.2/1,000 such deliveries, n5439), in which blood products had been transfused antenatally (187.5/1,000, n563) and when a postpartum hemorrhage occurred in the birth admission (135.3/1,000, n58,388). In women with no prior indication, the transfusion rate was 9.5 per 1,000 deliveries (n56,927). Among birth admissions with fewer than a 4-day antenatal stay involving transfusion, 80.8% included a hemorrhage diagnosis. Women who received transfusions in birth or postnatal admission were in the hospital longer than those who did not (median days [interquartile range] birth 5 [4, 5] compared with 3 [2, 5], postnatal 3 [2, 5] compared with 2 [0, 3]). Full data on possible confounders were available for 885,389 deliveries (99.3%) and were included in our risk factor analysis. Overall, 28.1% of women delivered by cesarean (n5249,775). Considering all births after adjusting for maternal factors, forceps delivery was associated with the highest risk of transfusion of all modes of birth when compared with noninstrumental vaginal delivery followed by intrapartum cesarean delivery, vacuum delivery, and prelabor cesarean delivery (Table 2). Across both vaginal and cesarean deliveries, women with bleeding or platelet disorders or placenta previa were at increased risk of transfusion. Among vaginal births, forceps deliveries and multiple births were also associated with more than doubling of the risk of transfusion, whereas among cesarean deliveries, preterm births at less than 33 weeks of gestation and antepartum hemorrhage were associated with a more than doubling of risk. Nulliparous women were at greater risk of transfusion when delivering vaginally as were women with a previous cesarean delivery (Table 2). When included in the model, iron deficiency anemia was associated with an increased risk of transfusion (vaginal: number transfused5344, relative risk 7.1 [6.2–8.2]; cesarean: n5258, relative risk 4.7 [3.7–5.8]), leaving other estimates largely unchanged. However, identification of women with iron deficiency anemia has low sensitivity (5.7–12.0%)21 and so this was excluded from the final model.

DISCUSSION Transfusion of blood or blood products occurred in one in every 71 deliveries in New South Wales between 2001 and 2010 with the majority of these related to hemorrhage. The majority of transfusions occurred during the birth admission, which is unsurprising given the large proportion associated with postpartum hem-

VOL. 123, NO. 1, JANUARY 2014

orrhage, which typically occurs within 24 hours after birth. The transfusion rate among women with no prior indication for blood transfusion was 1.0%, indicating that transfusion in otherwise healthy women is not uncommon. Over the last 10 years there has been a 33% increase in the use of blood and blood products throughout pregnancy. Obstetric patients use a small proportion of the blood supply overall (3–4%)22; however, there is potentially increasing demand in this subpopulation at a time when resources are becoming more limited.1 One possible explanation for the increase is the changing maternal population with more older mothers, women with previous cesarean deliveries, and having more comorbid conditions6,11; however, we found that the trend persisted when changes in these factors were taken into account. Similarly, Kuklina et al7 found the increasing trend in transfusion between 1998 and 2005 in the United States persisted despite adjustment for confounders. Another possible reason for the increase in transfusion is the recent increase in postpartum hemorrhage, which has been observed both in New South Wales23 and internationally.6,8,11,13 Because the majority of transfusions are performed in admissions with a diagnosis of hemorrhage, an increase in postpartum hemorrhage would likely lead to an increase in blood transfusions as has been observed elsewhere.8,9 The increase in transfusion is also possibly linked to increased severity of postpartum hemorrhage.9,11 The increase in transfusion rates may also reflect a change in practice with clinicians using less restrictive criteria for transfusion than in the past. We were unable to assess changes in transfusion thresholds; however, given the growing awareness of risks and costs associated with transfusion,1,2 it would be expected that changing practice would reduce transfusion rates. Interestingly, the slight dip in transfusion rates in 2008 coincided with a statewide initiative to decrease obstetric transfusions24; however, this decline was not maintained. Risk factors identified in this study were consistent with previous studies.10,16,25–27 Of note, vaginal birth after cesarean delivery carried a higher risk of transfusion than repeat cesarean delivery,10,16,27 and cesarean and instrumental delivery had a higher risk of transfusion than noninstrumental vaginal delivery.10,16,26 Higher rates of transfusion were found for preterm deliveries26 and are possibly related to anemia, which is a risk factor for both preterm birth and transfusion.28 We found the rate of transfusion to be much lower in private hospitals compared with public, which is consistent with lower-risk women giving

Patterson et al

Obstetric Blood Transfusion

131

birth in appropriate settings.29–31 In Australia, tertiary obstetric care is only available in public hospitals. Tertiary hospitals had the highest rates of transfusion of platelets and coagulation factors, which may reflect increased complexity of cases in these hospitals or better access to products in the larger centers. This study reflects the population burden of blood and blood product use in obstetrics in Australia. The use of longitudinally linked data allowed for the examination of blood product use within pregnancy. Validated and reliably collected information was available. The large number of women in the sample allowed for adjustment by a range of risk factors. Administrative data sets however lack clinical detail such as quantity of blood transfused, indication for transfusion, and hemoglobin measurements, which would provide insight into the severity of patients requiring transfusion. Obstetric transfusion represents a small proportion of overall blood use; however, use of blood in obstetrics is rising and there is potential for this to continue as postpartum hemorrhage rates continue to increase. Because it has not been possible in this study to determine the exact indications and “triggers” for red cell transfusions, it is difficult to opine as to the appropriateness of many of the transfusions. Clearly, in exsanguinating hemorrhage, transfusion is essential and a life-saving measure with questions revolving more around hemodynamic and coagulopathic parameters. On the other hand, in hemodynamically stable patients in whom hemorrhage has been controlled or is not the problem, the issues of patient blood management and transfusion are different. Some reduction in transfusion rates may be possible through increased awareness of transfusion risk factors and treatment of anemia during pregnancy and adherence to principles of patient blood management. Additional reduction in blood use may be achievable through exploring variation in blood use between hospitals. REFERENCES 1. Thomson A, Farmer S, Hofmann A, Isbister J, Shander A. Patient blood management—a new paradigm for transfusion medicine? ISBT Sci Ser 2009;4:423–35. 2. Goodnough LT, Shander A. Patient blood management. Anesthesiology 2012;116:1367–76.

5. Roberts CL, Ford JB, Algert CS, Bell JC, Simpson JM, Morris JM. Trends in adverse maternal outcomes during childbirth: a population-based study of severe maternal morbidity. BMC Pregnancy Childbirth 2009;9:7. 6. Knight M, Callaghan WM, Berg C, Alexander S, BouvierColle MH, Ford JB, et al. Trends in postpartum hemorrhage in high resource countries: a review and recommendations from the International Postpartum Hemorrhage Collaborative Group. BMC Pregnancy Childbirth 2009;9:55. 7. Kuklina EV, Meikle SF, Jamieson DJ, Whiteman MK, Barfield WD, Hillis SD, et al. Severe obstetric morbidity in the United States: 1998–2005. Obstet Gynecol 2009;113: 293–9. 8. Callaghan WM, Kuklina EV, Berg CJ. Trends in postpartum hemorrhage: United States, 1994–2006. Am J Obstet Gynecol 2010;202:353.e1–6. 9. Mehrabadi A, Hutcheon JA, Lee L, Liston RM, Joseph KS. Trends in postpartum hemorrhage from 2000 to 2009: a population-based study. BMC Pregnancy Childbirth 2012;12: 108. 10. Jakobsson M, Gissler M, Tapper AM. Risk factors for blood transfusion at delivery in Finland. Acta Obstet Gynecol Scand 2013;92:414–20. 11. Lutomski JE, Byrne BM, Devane D, Greene RA. Increasing trends in atonic postpartum haemorrhage in Ireland: an 11-year population-based cohort study. BJOG 2012;119:306–14. 12. Callaghan WM, Creanga AA, Kuklina EV. Severe maternal morbidity among delivery and postpartum hospitalizations in the United States. Obstet Gynecol 2012;120:1029–36. 13. Joseph KS, Rouleau J, Kramer MS, Young DC, Liston RM, Baskett TF; Maternal Health Study Group of the Canadian Perinatal Surveillance System. Investigation of an increase in postpartum haemorrhage in Canada. BJOG 2007;114:751–9. 14. McLintock C, James AH. Obstetric hemorrhage. J Thromb Haemost 2011;9:1441–51. 15. Al-Zirqi I, Vangen S, Forsen L, Stray-Pedersen B. Prevalence and risk factors of severe obstetric haemorrhage. BJOG 2008; 115:1265–72. 16. Jou HJ, Hung HW, Yan YH, Wu SC. Risk factors for blood transfusion in singleton pregnancy deliveries in Taiwan. Int J Gynaecol Obstet 2012;117:124–7. 17. Lain SJ, Roberts CL, Hadfield RM, Bell JC, Morris JM. How accurate is the reporting of obstetric haemorrhage in hospital discharge data? A validation study. Aust N Z J Obstet Gynaecol 2008;48:481–4. 18. Centre for Health Record Linkage. CHeReL—technical details. 2011. Available at: http://www.cherel.org.au/how-record-linkage-works/technical-details. Retrieved June 5, 2013. 19. National Centre for Classification in Health. Australian classification of health interventions. Sydney (Australia): National Centre for Classification in Health; 2006. 20. National Centre for Classification in Health. Australian Coding Standards for ICD-10-AM and ACHI. Sydney (Australia): National Centre for Classification in Health; 2006.

3. Shander A, Van Aken H, Colomina MJ, Gombotz H, Hofmann A, Krauspe R, et al. Patient blood management in Europe. Br J Anaesth 2012;109:55–68.

21. Lain SJ, Hadfield RM, Raynes-Greenow CH, Ford JB, Mealing NM, Algert CS, et al. Quality of data in perinatal population health databases: a systematic review. Med Care 2012;50:e7–20.

4. Sapere Research Group. Analysis of cost drivers and trends in the blood sector. Options manage appropriate use blood and blood products. Sydney (Australia): Sapere Research Group; 2011.

22. Allden RL, Sinha R, Roxby DJ, Ireland S, Hakendorf P, Robinson KL. Red alert—a new perspective on patterns of blood use in the South Australian public sector. Aust Health Rev 2011;35:327–33.

132

Patterson et al

Obstetric Blood Transfusion

OBSTETRICS & GYNECOLOGY

23. Ford JB, Roberts CL, Simpson JM, Vaughan J, Cameron CA. Increased postpartum hemorrhage rates in Australia. Int J Gynaecol Obstet 2007;98:237–43.

28. Scholl TO. Iron status during pregnancy: setting the stage for mother and infant. Am J Clin Nutr 2005;81: 1218S–22S.

24. Evaluation report for blood Watch Campaign 2009. 2010. Available at: http://www.cec.health.nsw.gov.au/__documents/programs/ blood-watch/303evaluation_2009.pdf. Retrieved June 5, 2013.

29. Olive EC, Roberts CL, Algert CS, Morris JM. Placenta praevia: maternal morbidity and place of birth. Aust N Z J Obstet Gynaecol 2005;45:499–504.

25. Huq FY, Kadir RA. Management of pregnancy, labour and delivery in women with inherited bleeding disorders. Haemophilia 2011;17(suppl 1):20–30.

30. Royal College of Obstetricians and Gynaecologists. Blood transfusion in obstetrics. Green-top Guideline No. 47. London (UK): Royal College of Obstetricians and Gynaecologists; 2007.

26. Ehrenthal DB, Chichester ML, Cole OS, Jiang X. Maternal risk factors for peripartum transfusion. J Womens Health (Larchmt) 2012;21:792–7. 27. Holm C, Langhoff-Roos J, Petersen KB, Norgaard A, Diness BR. Severe postpartum haemorrhage and mode of delivery: a retrospective cohort study. BJOG 2012;119:596–604.

31. The Royal Australian and New Zealand College of Obstetricians and Gynaecologists. Management of postpartum haemorrhage (PPH). Melbourne (Australia): The Royal Australian and New Zealand College of Obstetricians and Gynaecologists; 2011.

Build Your Reference List at www.greenjournal.org The Green Journal’s web site now makes research easier. The full-text view of each article features an “Article Tools” box. Clicking on the link “Export to CitationManager” allows you to export the citation information to EndNote®, ProCite®, or ReferenceManager®. rev 7/2013

VOL. 123, NO. 1, JANUARY 2014

Patterson et al

Obstetric Blood Transfusion

133