Association of Timing of Plasma Transfusion With Adverse Maternal Outcomes in Women With Persistent Postpartum Hemorrhage | Hematology | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Time-Dependent Propensity Score Matching of Women With Persistent Postpartum Hemorrhage (PPH)
Time-Dependent Propensity Score Matching of Women With Persistent Postpartum Hemorrhage (PPH)

Propensity score is the probability of plasma transfusion at a specific time point, given the woman’s observed characteristics at that time point.

Figure 2.  Derivation of Study Population
Derivation of Study Population

FFP indicates fresh frozen plasma; IQR, interquartile range; PPH, postpartum hemorrhage; and RBC, red blood cell.

Figure 3.  Balance of Covariate Values After Time-Dependent Propensity Score Matching of Women With Persistent Postpartum Hemorrhage
Balance of Covariate Values After Time-Dependent Propensity Score Matching of Women With Persistent Postpartum Hemorrhage

AIP indicates abnormally invasive placenta and RBC, red blood cell.

Table 1.  Characteristics of Women With Persistent PPH in the Total Cohort and the Propensity Score–Matched Cohort Stratified by Plasma Transfusion Strategy
Characteristics of Women With Persistent PPH in the Total Cohort and the Propensity Score–Matched Cohort Stratified by Plasma Transfusion Strategy
Table 2.  Outcomes of Women With Persistent PPH in the Total Cohort and the Propensity Score–Matched Cohort Stratified by Plasma Transfusion Strategy
Outcomes of Women With Persistent PPH in the Total Cohort and the Propensity Score–Matched Cohort Stratified by Plasma Transfusion Strategy
1.
Say  L, Chou  D, Gemmill  A,  et al.  Global causes of maternal death: a WHO systematic analysis.  Lancet Glob Health. 2014;2(6):e323-e333. doi:10.1016/S2214-109X(14)70227-XPubMedGoogle ScholarCrossref
2.
Zwart  JJ, Dupuis  JR, Richters  A, Ory  F, van Roosmalen  J.  Obstetric intensive care unit admission: a 2-year nationwide population-based cohort study.  Intensive Care Med. 2010;36(2):256-263. doi:10.1007/s00134-009-1707-xPubMedGoogle ScholarCrossref
3.
Grobman  WA, Bailit  JL, Rice  MM,  et al; Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units (MFMU) Network.  Frequency of and factors associated with severe maternal morbidity.  Obstet Gynecol. 2014;123(4):804-810. doi:10.1097/AOG.0000000000000173PubMedGoogle ScholarCrossref
4.
Al-Zirqi  I, Vangen  S, Forsen  L, Stray-Pedersen  B.  Prevalence and risk factors of severe obstetric haemorrhage.  BJOG. 2008;115(10):1265-1272. doi:10.1111/j.1471-0528.2008.01859.xPubMedGoogle ScholarCrossref
5.
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(6):751-759. doi:10.1111/j.1471-0528.2007.01316.xPubMedGoogle ScholarCrossref
6.
Mehrabadi  A, Liu  S, Bartholomew  S,  et al; Maternal Health Study Group of the Canadian Perinatal Surveillance System (Public Health Agency of Canada).  Temporal trends in postpartum hemorrhage and severe postpartum hemorrhage in Canada from 2003 to 2010.  J Obstet Gynaecol Can. 2014;36(1):21-33. doi:10.1016/S1701-2163(15)30680-0PubMedGoogle ScholarCrossref
7.
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. doi:10.1186/1471-2393-9-7PubMedGoogle ScholarCrossref
8.
Abdul-Kadir  R, McLintock  C, Ducloy  AS,  et al.  Evaluation and management of postpartum hemorrhage: consensus from an international expert panel.  Transfusion. 2014;54(7):1756-1768. doi:10.1111/trf.12550PubMedGoogle ScholarCrossref
9.
Bonnet  MP, Basso  O.  Prohemostatic interventions in obstetric hemorrhage.  Semin Thromb Hemost. 2012;38(3):259-264. doi:10.1055/s-0032-1302441PubMedGoogle ScholarCrossref
10.
James  AH, Grotegut  C, Ahmadzia  H, Peterson-Layne  C, Lockhart  E.  Management of Coagulopathy in Postpartum Hemorrhage.  Semin Thromb Hemost. 2016;42(7):724-731. doi:10.1055/s-0036-1593417PubMedGoogle ScholarCrossref
11.
James  AH, McLintock  C, Lockhart  E.  Postpartum hemorrhage: when uterotonics and sutures fail.  Am J Hematol. 2012;87(suppl 1):S16-S22. doi:10.1002/ajh.23156PubMedGoogle ScholarCrossref
12.
McLintock  C, James  AH.  Obstetric hemorrhage.  J Thromb Haemost. 2011;9(8):1441-1451. doi:10.1111/j.1538-7836.2011.04398.xPubMedGoogle ScholarCrossref
13.
Holcomb  JB, Tilley  BC, Baraniuk  S,  et al; PROPPR Study Group.  Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: the PROPPR randomized clinical trial.  JAMA. 2015;313(5):471-482. doi:10.1001/jama.2015.12PubMedGoogle ScholarCrossref
14.
Pacheco  LD, Saade  GR, Costantine  MM, Clark  SL, Hankins  GD.  An update on the use of massive transfusion protocols in obstetrics.  Am J Obstet Gynecol. 2016;214(3):340-344. doi:10.1016/j.ajog.2015.08.068PubMedGoogle ScholarCrossref
15.
Pasquier  P, Gayat  E, Rackelboom  T,  et al.  An observational study of the fresh frozen plasma: red blood cell ratio in postpartum hemorrhage.  Anesth Analg. 2013;116(1):155-161. doi:10.1213/ANE.0b013e31826f084dPubMedGoogle ScholarCrossref
16.
Saule  I, Hawkins  N.  Transfusion practice in major obstetric haemorrhage: lessons from trauma.  Int J Obstet Anesth. 2012;21(1):79-83. doi:10.1016/j.ijoa.2011.09.009PubMedGoogle ScholarCrossref
17.
Shields  LE, Smalarz  K, Reffigee  L, Mugg  S, Burdumy  TJ, Propst  M.  Comprehensive maternal hemorrhage protocols improve patient safety and reduce utilization of blood products.  Am J Obstet Gynecol. 2011;205(4):368.e1-368.e8. doi:10.1016/j.ajog.2011.06.084PubMedGoogle ScholarCrossref
18.
Shields  LE, Wiesner  S, Fulton  J, Pelletreau  B.  Comprehensive maternal hemorrhage protocols reduce the use of blood products and improve patient safety.  Am J Obstet Gynecol. 2015;212(3):272-280. doi:10.1016/j.ajog.2014.07.012PubMedGoogle ScholarCrossref
19.
Burtelow  M, Riley  E, Druzin  M, Fontaine  M, Viele  M, Goodnough  LT.  How we treat: management of life-threatening primary postpartum hemorrhage with a standardized massive transfusion protocol.  Transfusion. 2007;47(9):1564-1572. doi:10.1111/j.1537-2995.2007.01404.xPubMedGoogle ScholarCrossref
20.
Collins  PW, Cannings-John  R, Bruynseels  D,  et al; OBS2 study collaborators.  Viscoelastometry guided fresh frozen plasma infusion for postpartum haemorrhage: OBS2, an observational study.  Br J Anaesth. 2017;119(3):422-434. doi:10.1093/bja/aex245PubMedGoogle ScholarCrossref
21.
Netherlands Trial Register.  Transfusion strategies in women during major obstetric haemorrhage. https://www.trialregister.nl/trial/3909. Accessed July 19, 2019.
22.
Sentilhes  L, Vayssière  C, Deneux-Tharaux  C,  et al.  Postpartum hemorrhage: guidelines for clinical practice from the French College of Gynaecologists and Obstetricians (CNGOF): in collaboration with the French Society of Anesthesiology and Intensive Care (SFAR).  Eur J Obstet Gynecol Reprod Biol. 2016;198:12-21. doi:10.1016/j.ejogrb.2015.12.012PubMedGoogle ScholarCrossref
23.
Henriquez  DDCA, Bloemenkamp  KWM, Loeff  RM,  et al; TeMpOH-1 study group.  Fluid resuscitation during persistent postpartum haemorrhage and maternal outcome: a nationwide cohort study.  Eur J Obstet Gynecol Reprod Biol. 2019;235:49-56. doi:10.1016/j.ejogrb.2019.01.027PubMedGoogle ScholarCrossref
24.
Shakur  H, Roberts  I, Bautista  R,  et al; CRASH-2 trial collaborators.  Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial.  Lancet. 2010;376(9734):23-32. doi:10.1016/S0140-6736(10)60835-5PubMedGoogle ScholarCrossref
25.
WOMAN Trial Collaborators.  Effect of early tranexamic acid administration on mortality, hysterectomy, and other morbidities in women with post-partum haemorrhage (WOMAN): an international, randomised, double-blind, placebo-controlled trial.  Lancet. 2017;389(10084):2105-2116. doi:10.1016/S0140-6736(17)30638-4PubMedGoogle ScholarCrossref
26.
Andersen  LW, Granfeldt  A, Callaway  CW,  et al; American Heart Association’s Get With The Guidelines–Resuscitation Investigators.  Association between tracheal intubation during adult in-hospital cardiac arrest and survival.  JAMA. 2017;317(5):494-506. doi:10.1001/jama.2016.20165PubMedGoogle ScholarCrossref
27.
Andersen  LW, Raymond  TT, Berg  RA,  et al; American Heart Association’s Get With The Guidelines–Resuscitation Investigators.  Association between tracheal intubation during pediatric in-hospital cardiac arrest and survival.  JAMA. 2016;316(17):1786-1797. doi:10.1001/jama.2016.14486PubMedGoogle ScholarCrossref
28.
Austin  PC.  An introduction to propensity score methods for reducing the effects of confounding in observational studies.  Multivariate Behav Res. 2011;46(3):399-424. doi:10.1080/00273171.2011.568786PubMedGoogle ScholarCrossref
29.
Lu  B.  Propensity score matching with time-dependent covariates.  Biometrics. 2005;61(3):721-728. doi:10.1111/j.1541-0420.2005.00356.xPubMedGoogle ScholarCrossref
30.
Nakahara  S, Tomio  J, Takahashi  H,  et al.  Evaluation of pre-hospital administration of adrenaline (epinephrine) by emergency medical services for patients with out of hospital cardiac arrest in Japan: controlled propensity matched retrospective cohort study.  BMJ. 2013;347:f6829. doi:10.1136/bmj.f6829PubMedGoogle ScholarCrossref
31.
Xie  Y, Bowe  B, Li  T, Xian  H, Yan  Y, Al-Aly  Z.  Risk of death among users of proton pump inhibitors: a longitudinal observational cohort study of United States veterans.  BMJ Open. 2017;7(6):e015735. doi:10.1136/bmjopen-2016-015735PubMedGoogle Scholar
32.
Bolliger  D, Görlinger  K, Tanaka  KA.  Pathophysiology and treatment of coagulopathy in massive hemorrhage and hemodilution.  Anesthesiology. 2010;113(5):1205-1219. doi:10.1097/ALN.0b013e3181f22b5aPubMedGoogle ScholarCrossref
33.
Hancock  A, Weeks  AD, Lavender  DT.  Is accurate and reliable blood loss estimation the ‘crucial step’ in early detection of postpartum haemorrhage: an integrative review of the literature.  BMC Pregnancy Childbirth. 2015;15:230. doi:10.1186/s12884-015-0653-6PubMedGoogle ScholarCrossref
34.
Pallasmaa  N, Ekblad  U, Gissler  M, Alanen  A.  The impact of maternal obesity, age, pre-eclampsia and insulin dependent diabetes on severe maternal morbidity by mode of delivery: a register-based cohort study.  Arch Gynecol Obstet. 2015;291(2):311-318. doi:10.1007/s00404-014-3352-zPubMedGoogle ScholarCrossref
35.
Schols  SE, Lancé  MD, Feijge  MA,  et al.  Impaired thrombin generation and fibrin clot formation in patients with dilutional coagulopathy during major surgery.  Thromb Haemost. 2010;103(2):318-328. doi:10.1160/TH09-06-0396PubMedGoogle ScholarCrossref
36.
Henriquez  DDCA, Bloemenkamp  KWM, van der Bom  JG.  Management of postpartum hemorrhage: how to improve maternal outcomes [published online June 8, 2018].  J Thromb Haemost. doi:10.1111/jth.14200Google Scholar
37.
Austin  PC.  Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies.  Pharm Stat. 2011;10(2):150-161. doi:10.1002/pst.433PubMedGoogle ScholarCrossref
38.
Austin  PC.  A comparison of 12 algorithms for matching on the propensity score.  Stat Med. 2014;33(6):1057-1069. doi:10.1002/sim.6004PubMedGoogle ScholarCrossref
39.
Lunt  M.  Selecting an appropriate caliper can be essential for achieving good balance with propensity score matching.  Am J Epidemiol. 2014;179(2):226-235. doi:10.1093/aje/kwt212PubMedGoogle ScholarCrossref
40.
Rozé  JC, Cambonie  G, Marchand-Martin  L,  et al; Hemodynamic EPIPAGE 2 Study Group.  Association between early screening for patent ductus arteriosus and in-hospital mortality among extremely preterm infants.  JAMA. 2015;313(24):2441-2448. doi:10.1001/jama.2015.6734PubMedGoogle ScholarCrossref
41.
Janssen  KJ, Donders  AR, Harrell  FE  Jr,  et al.  Missing covariate data in medical research: to impute is better than to ignore.  J Clin Epidemiol. 2010;63(7):721-727. doi:10.1016/j.jclinepi.2009.12.008PubMedGoogle ScholarCrossref
42.
Moons  KG, Donders  RA, Stijnen  T, Harrell  FE  Jr.  Using the outcome for imputation of missing predictor values was preferred.  J Clin Epidemiol. 2006;59(10):1092-1101. doi:10.1016/j.jclinepi.2006.01.009PubMedGoogle ScholarCrossref
43.
Li  P, Stuart  EA, Allison  DB.  Multiple imputation: a flexible tool for handling missing data.  JAMA. 2015;314(18):1966-1967. doi:10.1001/jama.2015.15281PubMedGoogle ScholarCrossref
44.
Mitra  R, Reiter  JP.  A comparison of two methods of estimating propensity scores after multiple imputation.  Stat Methods Med Res. 2016;25(1):188-204. doi:10.1177/0962280212445945PubMedGoogle ScholarCrossref
45.
Penning de Vries  B, Groenwold  R.  Comments on propensity score matching following multiple imputation.  Stat Methods Med Res. 2016;25(6):3066-3068. doi:10.1177/0962280216674296PubMedGoogle ScholarCrossref
46.
Rubin  DB, Schenker  N.  Multiple imputation in health-care databases: an overview and some applications.  Stat Med. 1991;10(4):585-598. doi:10.1002/sim.4780100410PubMedGoogle ScholarCrossref
47.
Austin  PC.  Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples.  Stat Med. 2009;28(25):3083-3107. doi:10.1002/sim.3697PubMedGoogle ScholarCrossref
48.
Belitser  SV, Martens  EP, Pestman  WR, Groenwold  RH, de Boer  A, Klungel  OH.  Measuring balance and model selection in propensity score methods.  Pharmacoepidemiol Drug Saf. 2011;20(11):1115-1129. doi:10.1002/pds.2188PubMedGoogle ScholarCrossref
49.
Ali  MS, Groenwold  RH, Belitser  SV,  et al.  Reporting of covariate selection and balance assessment in propensity score analysis is suboptimal: a systematic review.  J Clin Epidemiol. 2015;68(2):112-121. doi:10.1016/j.jclinepi.2014.08.011PubMedGoogle ScholarCrossref
50.
Groenwold  RH, de Vries  F, de Boer  A,  et al.  Balance measures for propensity score methods: a clinical example on beta-agonist use and the risk of myocardial infarction.  Pharmacoepidemiol Drug Saf. 2011;20(11):1130-1137. doi:10.1002/pds.2251PubMedGoogle ScholarCrossref
51.
Normand  ST, Landrum  MB, Guadagnoli  E,  et al.  Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores.  J Clin Epidemiol. 2001;54(4):387-398. doi:10.1016/S0895-4356(00)00321-8PubMedGoogle ScholarCrossref
52.
Haukoos  JS, Lewis  RJ.  The propensity score.  JAMA. 2015;314(15):1637-1638. doi:10.1001/jama.2015.13480PubMedGoogle ScholarCrossref
53.
Collins  PW, Cannings-John  R, Bruynseels  D,  et al.  Viscoelastometric-guided early fibrinogen concentrate replacement during postpartum haemorrhage: OBS2, a double-blind randomized controlled trial.  Br J Anaesth. 2017;119(3):411-421. doi:10.1093/bja/aex181PubMedGoogle ScholarCrossref
54.
Wikkelsø  AJ, Edwards  HM, Afshari  A,  et al; FIB-PPH trial group.  Pre-emptive treatment with fibrinogen concentrate for postpartum haemorrhage: randomized controlled trial.  Br J Anaesth. 2015;114(4):623-633. doi:10.1093/bja/aeu444PubMedGoogle ScholarCrossref
55.
Gillissen  A, van den Akker  T, Caram-Deelder  C,  et al; TeMpOH-1 Study Group.  Coagulation parameters during the course of severe postpartum hemorrhage: a nationwide retrospective cohort study.  Blood Adv. 2018;2(19):2433-2442. doi:10.1182/bloodadvances.2018022632PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Original Investigation
    Obstetrics and Gynecology
    November 15, 2019

    Association of Timing of Plasma Transfusion With Adverse Maternal Outcomes in Women With Persistent Postpartum Hemorrhage

    Author Affiliations
    • 1Department of Obstetrics, Leiden University Medical Center, Leiden, the Netherlands
    • 2Jon J van Rood Center for Clinical Transfusion Research, Sanquin-Leiden University Medical Center, Leiden, the Netherlands
    • 3Department of Clinical Epidemiology, Leiden University Medical Center, Leiden, the Netherlands
    • 4Department of Biomedical Data Sciences, Leiden University Medical Center, Leiden, the Netherlands
    • 5Department of Obstetrics and Gynecology, Deventer Hospital, Deventer, the Netherlands
    • 6Athena Institute, Vrije Universiteit, Amsterdam, the Netherlands
    • 7Division of Thrombosis and Hemostasis, Department of Internal Medicine, Leiden University Medical Center, Leiden, the Netherlands
    • 8Sanquin Blood Bank, Unit Transfusion Medicine, Leiden, the Netherlands
    • 9Department of Immunohematology and Blood Transfusion, Leiden University Medical Center, Leiden, the Netherlands
    • 10Department of Anesthesiology, Albert Schweitzer Hospital, Dordrecht, the Netherlands
    • 11Department of Obstetrics, Birth Center Wilhelmina’s Children Hospital, Division Woman and Baby, University Medical Centre Utrecht, Utrecht, the Netherlands
    JAMA Netw Open. 2019;2(11):e1915628. doi:10.1001/jamanetworkopen.2019.15628
    Key Points español 中文 (chinese)

    Question  Is plasma transfusion within the first 60 minutes of persistent postpartum hemorrhage (PPH) associated with incidence of maternal adverse outcomes?

    Findings  In this cohort study of 114 propensity score–matched women with persistent PPH, plasma transfusion within the first 60 minutes of persistent PPH was not associated with incidence of maternal adverse outcomes compared with no or later plasma transfusion, independent of severity of PPH at the time of plasma transfusion.

    Meaning  These findings do not support the theory that early plasma transfusion in women with persistent PPH is better than no or later plasma transfusion.

    Abstract

    Importance  Early plasma transfusion for women with severe postpartum hemorrhage (PPH) is recommended to prevent coagulopathy. However, there is no comparative, quantitative evidence on the association of early plasma transfusion with maternal outcomes.

    Objective  To compare the incidence of adverse maternal outcomes among women who received plasma during the first 60 minutes of persistent PPH vs women who did not receive plasma for similarly severe persistent PPH.

    Design, Setting, and Participants  This multicenter cohort study used a consecutive sample of women with persistent PPH, defined as PPH refractory to first-line measures to control bleeding, between January 1, 2011, and January 1, 2013. Time-dependent propensity score matching was used to select women who received plasma during the first 60 minutes of persistent PPH and match each of them with a woman who had shown the same severity and received the same treatment of PPH but who had not received plasma at the moment of matching. Transfusions were not guided by coagulation tests. Statistical analysis was performed from June 2018 to June 2019.

    Exposures  Transfusion of plasma during the first 60 minutes of persistent PPH vs no or later plasma transfusion.

    Main Outcomes and Measures  Incidence of adverse maternal outcomes, defined as a composite of death, hysterectomy, or arterial embolization.

    Results  This study included 1216 women (mean [SD] age, 31.6 [5.0] years) with persistent PPH, of whom 932 (76.6%) delivered vaginally and 780 (64.1%) had PPH caused by uterine atony. Seven women (0.6%) died because of PPH, 62 women (5.1%) had a hysterectomy, and 159 women (13.1%) had arterial embolizations. Among women who received plasma during the first 60 minutes of persistent PPH, 114 women could be matched with a comparable woman who had not received plasma at the moment of matching. The incidence of adverse maternal outcomes was similar between the women, with adverse outcomes recorded in 24 women (21.2%) who received early plasma transfusion and 23 women (19.9%) who did not receive early plasma transfusion (odds ratio, 1.09; 95% CI, 0.57-2.09). Results of sensitivity analyses were comparable to the primary results.

    Conclusions and Relevance  In this cohort study, initiation of plasma transfusion during the first 60 minutes of persistent PPH was not associated with adverse maternal outcomes compared with no or later plasma transfusion, independent of severity of PPH.

    Introduction

    Obstetric hemorrhage accounts for 27% of all maternal deaths.1 In high-resource settings, maternal death due to postpartum hemorrhage (PPH) has become uncommon, but PPH remains an important cause of severe maternal morbidity.2-7

    Women with persistent PPH are at risk of developing coagulopathy due to depletion of coagulation factors and platelets.8-12 Coagulopathy can eventually lead to worse maternal outcomes. Timely transfusion of plasma may prevent coagulopathy and thereby improve maternal outcomes.

    Results from a 2015 study13 among patients with trauma suggest that formulaic plasma transfusion, comprising a fixed ratio of plasma to red blood cells (RBCs), is associated with better outcomes. Whether such transfusion strategies are also associated with better outcomes among women with persistent PPH is not clear. Some studies have suggested that early and aggressive plasma transfusion has a positive association with outcomes in women with PPH.14-19 However, a 2017 study20 suggested that women with persistent PPH have better outcomes when plasma transfusion is postponed or even avoided. Uncertainty about the outcomes associated with plasma transfusion among women with persistent PPH can lead to significant variation in clinical practice. This variation in practice, along with careful documentation of confounding factors, enables the use of routinely collected clinical data to compare outcomes among women treated according to different treatment strategies.

    The aim of this study was to assess whether early plasma transfusion is associated with improved maternal outcomes in women with persistent PPH. Our hypothesis was that initiation of plasma transfusion during the first 60 minutes of persistent PPH would be associated with fewer adverse maternal outcomes, defined as maternal death, hysterectomy, or arterial embolization compared with women who received no or later plasma transfusion.

    Methods

    Approval was obtained from the Medical Ethics Research Committee of the Leiden University Medical Center and from the institutional review board of each study center, and a waiver of informed consent was granted because the study used deidentified data. The study was registered in the Netherlands Trial Register21 and reported according to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline.

    Study Design and Population

    The transfusion strategies in women during major obstetric hemorrhage (TeMpOH-1) study21 was a multicenter, retrospective cohort study in the Netherlands that included consecutive women who had received 4 or more units of RBCs or a multicomponent blood transfusion within 24 hours after giving birth because of severe PPH (ie, ≥1000 mL blood loss) from January 1, 2011, to January 1, 2013. A multicomponent blood transfusion was defined as transfusion of at least 1 unit of packed RBCs in combination with plasma or platelet concentrates. We selected women from transfusion databases and birth registries in 61 participating hospitals.

    From this cohort, we identified women with persistent PPH, defined as PPH with at least 1000 mL of blood loss refractory to first-line interventions to control bleeding.8,22 First-line interventions depended on the cause of bleeding, as previously described (eTable 1 in the Supplement).23 We regarded the time of initiation of the first-line intervention to stop PPH as the moment of diagnosis of persistent PPH, under the assumption that refractoriness to first-line treatment would become evident shortly after initiation of this therapy. Women were followed up from onset until cessation of PPH.

    We excluded women with unknown timing of initiation of plasma transfusion. We also excluded women with initiation of plasma transfusion for any reason other than correcting coagulopathy secondary to PPH (ie, comorbidity).

    Data Collection

    Trained medical students and research nurses uniformly performed comprehensive health record reviews. From routinely collected medical information, we reconstructed the treatment course of every woman with PPH. We checked all data for completeness and inconsistencies and repeated on-site health record review as necessary. Data included comorbidity; mode of birth; primary cause of hemorrhage; consecutive estimates of blood loss and time of estimations; blood pressure, heart rate, and time of measurements; volume of crystalloids and colloids for fluid resuscitation; time of transfusions of packed RBCs, plasma, and platelets; and time of obstetric, radiological, and hemostatic interventions to stop bleeding.

    Fresh Frozen Plasma Transfusion

    Women with plasma transfusions received 1 or more units of fresh frozen plasma during the treatment of persistent PPH. Transfusion of plasma was not guided by coagulation tests. The time to plasma transfusion was defined as the interval between the moment of diagnosis of persistent PPH and administration of the first unit of plasma.

    Previous studies on hemostatic interventions to treat coagulopathy in pregnant and nonpregnant patients with major hemorrhage showed beneficial associations of these interventions when initiated early after the start of hemorrhage, specifically within 3 hours.24,25 Therefore, we examined the association of plasma transfusion during the first 60 minutes of persistent PPH with maternal outcomes.

    Outcome

    The outcome was the incidence of adverse maternal outcomes, defined as a composite of death, hysterectomy, or arterial embolization to control bleeding. The end of bleeding was defined as the time of the final recorded measurement of blood loss or the time of the last obstetric intervention to stop bleeding.

    In the Netherlands, uterine or internal iliac artery embolization is performed before resorting to hysterectomy, if the woman’s hemodynamic condition is stable enough to perform this procedure. During our study, 83.6% of the hospitals had this treatment modality available 24 hours per day, 7 days per week, and 92.5% of our study population gave birth in 1 of these hospitals. If a hospital does not have this treatment modality available, it is common practice to transfer the woman with PPH to a nearby hospital with embolization facilities. Embolization has almost completely substituted ligation of uterine or internal iliac arteries in the Netherlands, and in our study, ligation of arteries was performed in 0.8% of women with persistent PPH.

    Statistical Analysis

    Women with more severe PPH are more likely to receive early plasma transfusion, which confounds the association of early plasma transfusion with maternal outcomes. We used time-dependent propensity score matching to ensure that the contrasted groups were similar in terms of severity of hemorrhage and other treatments for PPH.26-31 First, we calculated the predicted probability to receive early plasma transfusion for all women in the cohort. Second, we selected pairs of women with the same probability for receiving plasma transfusion. These pairs consisted of one woman who received early plasma transfusion and another woman who did not. Third, we compared the matched groups.

    Propensity Scores

    The propensity score reflects the estimated probability of initiation of plasma transfusion in women with persistent PPH, given the observed characteristics of the women at the time of initiation of plasma transfusion.28,29 We calculated a propensity score for every woman with persistent PPH by using a multivariable Cox proportional hazards model. The outcome variable in this model was time to plasma transfusion, and the linear predictor at any given minute from diagnosing persistent PPH was used as the propensity score. In women with initiation of plasma transfusion before diagnosing persistent PPH (ie, women with placental abruption), we considered the time of diagnosing persistent PPH as the time of initiating plasma transfusion.

    We included baseline and time-dependent covariates associated with initiation of plasma transfusion and maternal outcome in a Cox model to calculate propensity scores. Selection of these potentially confounding variables was based on clinical reasoning and prior knowledge.7,8,32-36 The baseline covariates were mode of birth (ie, vaginal or cesarean), cause of hemorrhage (ie, uterine atony, retained placenta, abnormally invasive placenta, or other), preeclampsia (yes or no), and volume of crystalloids and colloids for fluid resuscitation (continuous variable). We included the following time-dependent variables: estimated volume of blood loss (continuous variable), bleeding rate (continuous variable), hemorrhagic shock (yes or no), oxytocin infusion (yes or no), misoprostol (yes or no), ergometrine (yes or no), the prostaglandin E2 analogue sulprostone (yes or no), manual removal of placenta (yes or no), exploration of uterine cavity and genital tract with anesthesia (yes or no), intrauterine balloon tamponade (yes or no), tranexamic acid (yes or no), fibrinogen concentrate (yes or no), recombinant factor VIIa (yes or no), packed RBCs transfusion (categorized as 0, 1, 2, 3, or ≥4 units), and platelet transfusion (yes or no). Additional information on handling of the time-dependent covariates in statistical analyses is provided in eTable 2 in the Supplement.

    Matching

    We applied a 1:1 nearest-neighbor risk-set matching algorithm on the propensity score without replacement, with a maximum caliper width of 0.1 of the SD of the logit of the propensity score.37-40 In this way, we sequentially matched every woman with persistent PPH in whom plasma transfusion was initiated at any given time point (0-60 minutes after diagnosis of persistent PPH) to a woman with similar propensity score in whom plasma transfusion was not initiated before or at that same time point (Figure 1). In this matched counterpart, plasma transfusion may have been initiated at a later time during PPH. After cessation of PPH or after reaching an endpoint (ie, arterial embolization, hysterectomy, or death), a woman was no longer considered at risk for plasma transfusion for correction of coagulopathy during ongoing hemorrhage.

    Missing covariate data were imputed by using multiple imputation.41-43 We included all confounding variables, outcome variables, and parameters associated with the missing variables as predictive variables in the imputation models and generated 10 imputed data sets. We tested our Cox model for nonproportional hazards by adding interactions with time.

    In each imputed data set, we estimated the propensity score for initiation of plasma transfusion for each woman with persistent PPH. We performed a time-dependent propensity score matching within each of these imputed data sets, and then we pooled the effect estimates by averaging them according to the Rubin’s rule.44-46

    After matching, we performed a check of the balance between the confounding variables to ensure our propensity score model was specified correctly. To this end, we calculated the standardized differences in the confounding variables between the women with plasma transfusion during the first 60 minutes of persistent PPH and the women with no or later plasma transfusion in our matched cohort.47-50 Absolute standardized differences less than 10% are generally considered a good balance of the observed confounding variables.28,51,52

    Main and Sensitivity Analyses

    We used logistic regression to assess the adjusted association of plasma transfusion during the first 60 minutes of persistent PPH with adverse maternal outcomes; the composite maternal outcome was the dependent variable, and time of plasma transfusion (ie, early vs no or later transfusion) was the independent variable. We used robust SEs to calculate 95% CIs.

    We performed several sensitivity analyses to assess the robustness of our results and to assess whether our effect estimate was influenced by women with plasma at a later time point in our comparison group. First, we performed sensitivity analyses with initiation of plasma transfusion during the first 120 and 180 minutes of persistent PPH because a potential beneficial effect of correction of coagulopathy has been previously described within the first 3 hours after the onset of hemorrhage in obstetric and nonobstetric populations.24,25

    Second, we performed sensitivity analyses by excluding pairs of women with a crossover of the woman initially without plasma to treatment with plasma shortly after matching. These analyses were performed with a restriction of 15, 30, 45, and 60 minutes on the time interval of switching from no plasma to plasma treatment. For example, if a woman treated with plasma at 50 minutes was matched to a women without plasma until 50 minutes but with initiation of plasma at 64 minutes, we excluded this pair in the sensitivity analysis for no crossover within 15 minutes.

    Third, we performed sensitivity analyses by excluding pairs of women with a crossover of the woman initially without plasma to treatment with plasma while still being within the first 60 minutes of persistent PPH. For example, if a woman treated with plasma at 30 minutes was matched to a woman without plasma at 30 minutes but with initiation of plasma at 55 minutes, we excluded the pair from this sensitivity analysis.

    Results
    Population

    The cohort included 1391 women with PPH who received 4 or more units of packed RBCs or a multicomponent blood transfusion within 24 hours after birth (Figure 2). Of these women, we classified 1260 (90.6%) as having persistent PPH. We excluded 43 women with persistent PPH because of unknown time of initiation of plasma transfusion and 1 woman in whom plasma transfusion had been started before birth because of leukemia instead of obstetric hemorrhage. Our final cohort included 1216 women (mean [SD] age, 31.6 [5.0] years). Seven women (0.6%) died because of PPH, 62 women (5.1%) had a hysterectomy, and 159 women (13.1%) had arterial embolizations.

    A total of 598 women (49.2%) received plasma during ongoing PPH. Among women in the no or later plasma transfusion group, 618 women (57.1%) did not receive plasma and 465 women (42.9%) received plasma at a later time after matching. Median (interquartile range [IQR]) time to initiation of plasma transfusion was 105 (65-196) minutes. Overall, plasma transfusion was initiated during the first 60 minutes of persistent PPH in 133 women (10.9%), during the first 120 minutes in 338 women (27.8%), and during the first 180 minutes in 433 women (35.6%).

    Baseline and time-dependent characteristics of women with early plasma transfusion vs no or later plasma transfusion are presented in Table 1. We imputed missing data on volume of fluid resuscitation (16.0%) and hemorrhagic shock at moment of diagnosing persistent PPH (34.9%). For this latter time-dependent confounding variable, more data (ie, measured blood pressures and heart rates) became available for an increasing proportion of women with progression of the PPH. An adverse maternal outcome was observed in 30 women (22.6%) with plasma transfusion during the first 60 minutes of persistent PPH and in 175 women (16.2%) with no or later plasma transfusions (odds ratio, 1.51; 95% CI, 0.98-2.34) (Table 2).

    Time-Dependent Propensity Score–Matched Population

    The number of matched pairs of women with plasma transfusion during the first 60 minutes and women with no or later plasma transfusion fluctuated across the 10 imputed data sets. We found a pooled average of 114 matches of women with plasma transfusion during the first 60 minutes and women with no plasma or plasma transfusion at a later time during persistent PPH. Nineteen women with plasma transfusion during the first 60 minutes had no match on propensity score (Table 1). Median (IQR) time to plasma transfusion in women with plasma transfusion during the first 60 minutes was 40 (16-50) minutes. Of their matched counterparts, 47 women (41.2%) did not receive plasma during PPH and 67 women (58.8%) received plasma at a later time during PPH, with a median (IQR) time to plasma transfusion of 66 (47-90) minutes in these 67 women. Across the 10 imputed data sets, we included a pooled average of 29 women twice in this matched cohort: first as a woman with no or later plasma transfusion and later as a woman with plasma transfusion during the first 60 minutes.

    Outcomes in Adjusted Analyses

    The distribution of baseline and time-dependent covariates in the matched cohort were well balanced between women with plasma transfusion during the first 60 minutes and women with no or later plasma transfusion (Figure 3 and Table 2). In the matched cohort, we observed a pooled average of 24 adverse maternal outcomes (21.2%) in women with plasma transfusion within 60 minutes vs 23 adverse maternal outcomes (19.9%) in women with no or later plasma transfusion (odds ratio, 1.09; 95% CI, 0.57-2.09).

    Sensitivity Analyses

    Unadjusted and adjusted sensitivity analyses in women with plasma transfusion within 120 minutes and within 180 minutes vs no or later plasma transfusion within these intervals yielded similar results as the primary analysis (Table 2) (eTable 3 and eTable 4 in the Supplement). In the sensitivity analyses excluding pairs of women in which a woman crossed over from no or later plasma to plasma transfusion 15, 30, 45, or 60 minutes after matching, we also found effect estimates comparable to our main analysis (eTable 5 in the Supplement). In the sensitivity analysis excluding 29 pairs of women because of crossover from no or later plasma to plasma transfusion during the first 60 minutes of persistent PPH, the odds ratio was 0.94 (95% CI, 0.43-2.06) for the remaining pairs of women.

    Discussion

    In this multicenter, time-dependent propensity score–matched cohort study of women with persistent PPH, empirical, early plasma transfusion was not associated with better maternal outcomes compared with women who received no or later plasma transfusion. Similar results were observed in all sensitivity analyses.

    Early plasma transfusion is believed to improve maternal outcomes because it could prevent or treat coagulopathy occurring among women treated for persistent PPH. Studies evaluating the effect of plasma transfusion on outcomes of women with severe PPH are scarce, to our knowledge. Contrary to our findings, a single-center observational study15 among 142 women with severe PPH reported a decreased rate of advanced interventions associated with a high ratio of plasma to packed RBCs. In that study, only 41 women received plasma in the management of PPH. Similarly, high ratios of plasma to packed RBCs have been reported to improve maternal outcomes when incorporated within PPH protocols, but whether this improvement could be attributed to the transfusion strategy or to other parts of the protocol is unclear.17,18

    The observed absence of an effect of early plasma transfusion on maternal outcomes among women with persistent PPH may have several explanations. First, there may have been too few women who developed significant coagulopathy and therefore there was no need to treat or prevent it. This explanation is consistent with findings from studies among women with severe PPH in whom fibrinogen concentrate was administered early during hemorrhage to prevent and correct coagulopathy.53,54 In these studies, most women had not developed coagulopathy at the time of administration of fibrinogen, and outcomes did not improve. Yet, in the TeMpOH-1 study cohort,21 26% of women eventually reached a fibrinogen level of less than 200 mg/dL (to convert to micromoles per liter, multiply by 0.0294), and 5% of women reached this level after losing less than 2 L of blood,55 which suggests that the absence of coagulopathy in our cohort is not an explanation for our findings.

    Second, plasma might not be effective in preventing or treating coagulopathy in women with persistent PPH, or the dose of plasma may have been too low to show a difference. It is conceivable that personalized supplementation of factor concentrates would be a better strategy to prevent adverse outcomes among women with PPH.

    Third, 42.9% of the women in the control group were eventually also treated with plasma. Some of these women received plasma relatively shortly after the moment at which they had been matched to their rapidly treated counterpart. If such later plasma was as effective as early administration of plasma, that could explain the observed absence of association of early plasma transfusion with outcomes. Yet, sensitivity analysis among matched pairs without this problem showed similar results, suggesting that this also did not explain our findings.

    Limitations and Strengths

    Our findings had some limitations and should be interpreted with caution, as they may also be explained by residual confounding. Women with more severe PPH are more likely to be rapidly treated with plasma than women with less severe hemorrhages. Time-dependent propensity score matching permitted us to balance all measured prognostic factors at any time during PPH, but this technique does not account for the distribution of unknown or unmeasured confounders. Yet, the professionals treating the women with severe PPH in our cohort carefully documented all parameters that are generally considered relevant with respect to the severity and treatment of PPH, to our knowledge. We could not think of any other parameters that might explain the observed absence of association. In addition, our findings may also be explained by random error. The confidence interval around the point estimate included values between 0.57 and 2.09, suggesting that there may be a protective or harmful association of early plasma transfusion with maternal outcomes, in line with the findings of previous studies.15-20

    A strength of our study was the use of persistent PPH, an intuitive and pragmatic definition of severe PPH with easy translation to daily clinical practice, to select women for this analysis.8,22,36 In the Netherlands, clinical parameters and the times of interventions are carefully recorded during obstetric emergencies. Thus, we were able to make a detailed reconstruction of the course of PPH, and we had no loss to follow up. In addition, extensive sensitivity analyses showed consistent results.

    Conclusions

    This cohort study found that among women with persistent PPH, empirical early plasma transfusion was not associated with maternal deaths, hysterectomies, or arterial embolizations compared with no or later plasma transfusion. Results were carefully adjusted for severity of PPH and time-dependent confounding, but residual confounding cannot be ruled out because of the observational nature of the study design.

    Our findings do not suggest that plasma transfusion has no place in the treatment of women with severe PPH. Rather, our study underlines the importance of developing tools to diagnose coagulopathy during persistent PPH. These tools may enable individualization of treatment of women with persistent PPH by identifying women who develop coagulopathy during persistent PPH.

    Back to top
    Article Information

    Accepted for Publication: September 29, 2019.

    Published: November 15, 2019. doi:10.1001/jamanetworkopen.2019.15628

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2019 Henriquez DDCA et al. JAMA Network Open.

    Corresponding Author: Dacia D. C. A. Henriquez, MD, Jon J van Rood Center for Clinical Transfusion Research, Sanquin-Leiden University Medical Center, Plesmanlaan 1a, 2333 BZ Leiden, the Netherlands (d.d.c.a.henriquez@lumc.nl).

    Author Contributions: Drs Henriquez and Caram-Deelder had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Concept and design: Henriquez, le Cessie, Zwart, van Roosmalen, Eikenboom, So-Osman, van de Watering, Zwaginga, Bloemenkamp, van der Bom.

    Acquisition, analysis, or interpretation of data: All authors.

    Drafting of the manuscript: Henriquez, Caram-Deelder, van der Bom.

    Critical revision of the manuscript for important intellectual content: All authors.

    Statistical analysis: Henriquez, Caram-Deelder, le Cessie, van der Bom.

    Obtained funding: van der Bom.

    Administrative, technical, or material support: Henriquez, Caram-Deelder, van der Bom.

    Supervision: van der Bom.

    Conflict of Interest Disclosures: Dr Eikenboom reported receiving grants from CSL Behring and fees from Roche and Celgene (paid to Leiden University Medical Center) outside the submitted work. No other disclosures were reported.

    Funding/Support: This work was supported by a grant from Sanquin Research (grant No. PPOC-11-032).

    Role of the Funder/Sponsor: The funder had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    TeMpOH-1 Research Group: Cock M. C. Bank, MSc, and Yvette S. Snuif–de Lange, MD, PhD (Admiraal de Ruyter Hospital, Goes, the Netherlands); Adriaan J. van Gammeren, MSc, PhD, and Dimitri N. M. Papatsonis, MD, PhD (Amphia Hospital, Breda, the Netherlands); Harriet Klinkspoor, MSc, PhD, and Marjolein Kok, MD, PhD (Amsterdam Medical Centre, Amsterdam, the Netherlands); Bauke A. de Boer, MSc, PhD (Atalmedial, Leiden, the Netherlands); Josje Langenveld, MD, PhD, and Mathie P. G. Leers, MSc, PhD (Atrium Medical Centre, Heerlen, the Netherlands); Jart H. C. Diris, MSc, PhD, and Rene D. Kok, MD, PhD (Bernhoven Hospital, Uden, the Netherlands); Paula Engbers, MSc, PhD, and Marjo J. C. P. Hanssen, MD, PhD (Bethesda Hospital, Hoogeveen, the Netherlands); Wim J. van Wijngaarden, MD, PhD (Bronovo Hospital, Den Haag, the Netherlands); Daniela H. Schippers, MD, PhD, and Jos W. J. van der Stappen, MSc, PhD (Canisius-Wilhelmina Hospital, Nijmegen, the Netherlands); Tom H. M. Hasaart, MD, PhD, and Daan H. van de Kerkhof, MSc, PhD (Catharina Hospital, Eindhoven, the Netherlands); Jacques B. de Kok, MSc, PhD (Deventer Hospital, Deventer, the Netherlands); Gijs A. van Unnik, MD, PhD (Diaconessen Hospital, Leiden, the Netherlands); Wouter Kortlandt, MSc, PhD, and Nico W. E. Schuitemaker, MD, PhD (Diakonessen Hospital, Utrecht, the Netherlands); Friso M. C. Delemarre, MD, PhD, and Hans L. P. van Duijnhoven, MSc, PhD (Elkerliek Hospital, Helmond, the Netherlands); Hans J. Duvekot, MD, PhD (Erasmus Medical Centre, Rotterdam, the Netherlands); Sietske Hogenboom, MSc, PhD, and Gunilla Kleiverda, MD, PhD (Flevo Hospital, Almere, the Netherlands); Mylene J. W. van Etten–van Hulst, MD, PhD, and Karin Pagano Mirani–Oostdijk, MSc, PhD (Fransiscus Hospital, Roosendaal, the Netherlands); Corine van Kampen, MSc, PhD (Gelderse Vallei Hospital, Ede, the Netherlands); Martin J. N. Weinans, MD, PhD (Gelderse Vallei Hospital, Apeldoorn, the Netherlands); Henk J. Adriaanse, MSc, PhD, and Anjoke J. M. Huisjes, MD, PhD (Gelre Hospital, Apeldoorn, the Netherlands); Marieke A. M. Frasa, MSc, PhD, Jeffrey F. W. Keuren, MSc, PhD, and Claudia A. van Meir, MD, PhD (Groene Hart Hospital, Gouda, the Netherlands); Hanneke Feitsma, MD, PhD, and Francisca Hudig, MSc, PhD (Haga Hospital, Den Haag, the Netherlands); J. Marko Sikkema, MD, PhD (Hospital Group Twente, Almelo, the Netherlands); Marleen I. Baas, MD, PhD (Hospital Rivierenland Tiel, Tiel, the Netherlands); Michael A. Fouraux, MSc, PhD, and Godfried C. H Metz, MD, PhD (Ikazia Hospital, Rotterdam, the Netherlands); Bas W. H. Nij Bijvank, MD, PhD, and Henk J. M. Rondeel, MSc, PhD (Isala Clinics, Zwolle, the Netherlands); Jos M. T. Roelofsen, MD, PhD (Lange Land Hospital, Zoetermeer, the Netherlands); Marjolein Doesburg–van Kleffens, MSc, PhD, and Sander C. de Wit, MD, PhD (Maas Hospital Pantein, Boxmeer, the Netherlands); Hans Versendaal, MD, and Floor Weerkamp, MSc, PhD (Maasstad Hospital, Rotterdam, the Netherlands); Yvonne M. C. Henskens, MSc, PhD, and Liesbeth H. C. J. Scheepers, MD, PhD (Maastricht University Medical Centre, Maastricht, the Netherlands); David P. van der Ham, MD, PhD, and Jan W. Smit, MSc, PhD (Martini Hospital, Groningen, the Netherlands); Fedde van der Graaf, MSc, PhD, and Martina M. Porath, MD, PhD (Máxima Medical Centre, Veldhoven, the Netherlands); Paulien C. M. van der Salm, MD, PhD, and Merel van Wijnen, MSc, PhD (Meander Medical Centre, Amersfoort, the Netherlands); Martina Pontesilli, MD (Medical Centre Alkmaar, Alkmaar, the Netherlands); Frederique M. van Dunné, MD, PhD, and Gabrielle A. E. Ponjee, MSc, PhD (Medical Centre Haaglanden, Den Haag, the Netherlands); Marinka S. Post, MD, PhD, and Betty S. van der Veen, MSc, PhD (Medical Centre Leeuwarden, Leeuwarden, the Netherlands); Josien T. J. Brons, MD, PhD, and Jennita Slomp, MSc, PhD (Medical Centre Twente, Enschede, the Netherlands); Arjan de Mare, MSc, PhD (Medlon, Almelo, the Netherlands); Anja Leyte, MSc, PhD, and Eline S. A. van den Akker, MD, PhD (Onze Lieve Vrouwe Hospital, Amsterdam, the Netherlands); Henri de Wet, MD, PhD (Refaja Hospital Stadskanaal, Stadskanaal, the Netherlands); Diana M. R. van der Borden, MD, PhD (Regional Hospital Koningin Beatrix, Winterswijk, the Netherlands); Henk A. Bremer, MD, PhD, and G. H. Martine Tax, MD, PhD (Reinier de Graaf Hospital, Delft, the Netherlands); Marjolein J. de Vries, MD, PhD (Rijnland Hospital, Leiderdorp, the Netherlands); Karin de Boer, MD, PhD, and Harm de Waard, MSc, PhD (Rijnstate Hospital, Arnhem, the Netherlands); Ries H. de Keijzer, MSc, PhD (Rivierenland Tiel Hospital, Tiel, the Netherlands); Jan M. Burggraaff, MD, PhD, and Jos G. J. Pouwels, MSc, PhD (Scheper Hospital, Emmen, the Netherlands); Nicolette van Gemund, MD, PhD, and Lenneke Prinzen, MSc, PhD (Sint Franciscus Hospital, Rotterdam, the Netherlands); Henriet A. Hendriks, MSc, and Brenda B. J. Hermsen, MD, PhD (Sint Lucas Andreas Hospital, Amsterdam, the Netherlands); Stephan G. A. Koehorst, MSc, PhD, and Tamara E. M. Verhagen, MD, PhD (Slingeland Hospital, Doetinchem, the Netherlands); Erik van Beek, MD, PhD, and Chris M. Hackeng, MSc, PhD (St Antonius Hospital, Nieuwegein, the Netherlands); Peter J. Kabel, MSc, PhD, and Pieternel Steures, MD, PhD (St Elisabeth Hospital, Tilburg, the Netherlands); Ivo M. A. van Dooren, MD, PhD, and Etienne C. H. J. Michielse, MSc, PhD (St Jans Hospital, Weert, the Netherlands); Helena Chon, MSc, PhD, Marco Treskes, MSc, PhD, and Harry Visser, MD, PhD (Tergooi Hospital, Blaricum, the Netherlands); Esther Oostenveld, MD, PhD, and Dick H. M. Peters, MD (Tjongerschans Hospital, Heerenveen, the Netherlands); Maureen T. M. Franssen, MD, PhD, and J. Henk Meekers (University Medical Centre Groningen, Groningen, the Netherlands); Mallory D. Woiski, MD, PhD, and Liesbeth C. M. van Pampus, MSc, PhD (University Medical Centre St Radboud, Nijmegen, the Netherlands); Martijn A. Oudijk, MD, PhD, and Karen M. K. de Vooght, MSc, PhD (University Medical Centre Utrecht, Utrecht, the Netherlands); Rolf L. M. Cikot, MD, PhD, and Leendert J. Mostert, MSc, PhD (Van Weel-Bethesda Hospital, Dirksland, the Netherlands); Huib Ceelie, MSc, PhD, and Annemarie M. G. Huijssoon, MD, PhD (Vlietland Hospital, Schiedam, the Netherlands); Christianne J. M. de Groot, MD, PhD, and Otto Visser, MD, PhD (VU Medical Centre, Amsterdam, the Netherlands); Niels Jonker, MSc, PhD, and Annemarieke Koops, MD, PhD (Wilhelmina Hospital, Assen, the Netherlands); Angelo Hooker, MD, PhD, and Nada Osmanovic, MSc, PhD (Zaans Medical Centre, Zaandam, the Netherlands); Herman J. L. M. Ulenkate, MSc, PhD, and Bianca Visschers, MD, PhD (Zorgsaam Zeeuws-Vlaanderen Hospital, Terneuzen, the Netherlands); and Jeffrey F. W. Keuren, MSc, PhD, and Gratia D. M. Martens, MD, PhD (Zuwe Hofpoort Hospital, Woerden, the Netherlands).

    Additional Contributions: Rosanne M. Loeff, MSc, Renske J. Van Goeverden, MSc, Bram Eijlers, MSc, Anne Hillebrand, MSc, Saskia E. Spelmink, MD, Tom J. Beunder, MSc, Valerie Harskamp, MSc, Merlijn Wind, MSc, Mark D. Koning, MSc, Roos A. Cramer, MSc, Artyom Veenstra, MSc, Sharissa M. Smith, MSc, and Elmara E. Ensing, MSc, Clara Kolster-Bijdevaate, and Marjolein S. Bourgonje-Verhart (Leiden University Medical Center, Leiden, the Netherlands), Cecilia E. Bleeker-Taborh, and Ellen Roos-van Milligen (Sanquin Research Leiden, Leiden, the Netherlands) contributed to data collection. Karin J. Van Brussel-de Groot and Ouisam Zouitni, MSc (Sanquin Research Leiden, Leiden, the Netherlands), contributed to data cleaning. Rolf H. H. Groenwold, MD, PhD (Leiden University Medical Center, Leiden, the Netherlands), helped with the statistical analyses. Rachel E. Collis (Cardiff and Vale University, Cardiff, United Kingdom) provided helpful comments for improvement of this article. None of these individuals were compensated for their contribution.

    References
    1.
    Say  L, Chou  D, Gemmill  A,  et al.  Global causes of maternal death: a WHO systematic analysis.  Lancet Glob Health. 2014;2(6):e323-e333. doi:10.1016/S2214-109X(14)70227-XPubMedGoogle ScholarCrossref
    2.
    Zwart  JJ, Dupuis  JR, Richters  A, Ory  F, van Roosmalen  J.  Obstetric intensive care unit admission: a 2-year nationwide population-based cohort study.  Intensive Care Med. 2010;36(2):256-263. doi:10.1007/s00134-009-1707-xPubMedGoogle ScholarCrossref
    3.
    Grobman  WA, Bailit  JL, Rice  MM,  et al; Eunice Kennedy Shriver National Institute of Child Health and Human Development (NICHD) Maternal-Fetal Medicine Units (MFMU) Network.  Frequency of and factors associated with severe maternal morbidity.  Obstet Gynecol. 2014;123(4):804-810. doi:10.1097/AOG.0000000000000173PubMedGoogle ScholarCrossref
    4.
    Al-Zirqi  I, Vangen  S, Forsen  L, Stray-Pedersen  B.  Prevalence and risk factors of severe obstetric haemorrhage.  BJOG. 2008;115(10):1265-1272. doi:10.1111/j.1471-0528.2008.01859.xPubMedGoogle ScholarCrossref
    5.
    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(6):751-759. doi:10.1111/j.1471-0528.2007.01316.xPubMedGoogle ScholarCrossref
    6.
    Mehrabadi  A, Liu  S, Bartholomew  S,  et al; Maternal Health Study Group of the Canadian Perinatal Surveillance System (Public Health Agency of Canada).  Temporal trends in postpartum hemorrhage and severe postpartum hemorrhage in Canada from 2003 to 2010.  J Obstet Gynaecol Can. 2014;36(1):21-33. doi:10.1016/S1701-2163(15)30680-0PubMedGoogle ScholarCrossref
    7.
    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. doi:10.1186/1471-2393-9-7PubMedGoogle ScholarCrossref
    8.
    Abdul-Kadir  R, McLintock  C, Ducloy  AS,  et al.  Evaluation and management of postpartum hemorrhage: consensus from an international expert panel.  Transfusion. 2014;54(7):1756-1768. doi:10.1111/trf.12550PubMedGoogle ScholarCrossref
    9.
    Bonnet  MP, Basso  O.  Prohemostatic interventions in obstetric hemorrhage.  Semin Thromb Hemost. 2012;38(3):259-264. doi:10.1055/s-0032-1302441PubMedGoogle ScholarCrossref
    10.
    James  AH, Grotegut  C, Ahmadzia  H, Peterson-Layne  C, Lockhart  E.  Management of Coagulopathy in Postpartum Hemorrhage.  Semin Thromb Hemost. 2016;42(7):724-731. doi:10.1055/s-0036-1593417PubMedGoogle ScholarCrossref
    11.
    James  AH, McLintock  C, Lockhart  E.  Postpartum hemorrhage: when uterotonics and sutures fail.  Am J Hematol. 2012;87(suppl 1):S16-S22. doi:10.1002/ajh.23156PubMedGoogle ScholarCrossref
    12.
    McLintock  C, James  AH.  Obstetric hemorrhage.  J Thromb Haemost. 2011;9(8):1441-1451. doi:10.1111/j.1538-7836.2011.04398.xPubMedGoogle ScholarCrossref
    13.
    Holcomb  JB, Tilley  BC, Baraniuk  S,  et al; PROPPR Study Group.  Transfusion of plasma, platelets, and red blood cells in a 1:1:1 vs a 1:1:2 ratio and mortality in patients with severe trauma: the PROPPR randomized clinical trial.  JAMA. 2015;313(5):471-482. doi:10.1001/jama.2015.12PubMedGoogle ScholarCrossref
    14.
    Pacheco  LD, Saade  GR, Costantine  MM, Clark  SL, Hankins  GD.  An update on the use of massive transfusion protocols in obstetrics.  Am J Obstet Gynecol. 2016;214(3):340-344. doi:10.1016/j.ajog.2015.08.068PubMedGoogle ScholarCrossref
    15.
    Pasquier  P, Gayat  E, Rackelboom  T,  et al.  An observational study of the fresh frozen plasma: red blood cell ratio in postpartum hemorrhage.  Anesth Analg. 2013;116(1):155-161. doi:10.1213/ANE.0b013e31826f084dPubMedGoogle ScholarCrossref
    16.
    Saule  I, Hawkins  N.  Transfusion practice in major obstetric haemorrhage: lessons from trauma.  Int J Obstet Anesth. 2012;21(1):79-83. doi:10.1016/j.ijoa.2011.09.009PubMedGoogle ScholarCrossref
    17.
    Shields  LE, Smalarz  K, Reffigee  L, Mugg  S, Burdumy  TJ, Propst  M.  Comprehensive maternal hemorrhage protocols improve patient safety and reduce utilization of blood products.  Am J Obstet Gynecol. 2011;205(4):368.e1-368.e8. doi:10.1016/j.ajog.2011.06.084PubMedGoogle ScholarCrossref
    18.
    Shields  LE, Wiesner  S, Fulton  J, Pelletreau  B.  Comprehensive maternal hemorrhage protocols reduce the use of blood products and improve patient safety.  Am J Obstet Gynecol. 2015;212(3):272-280. doi:10.1016/j.ajog.2014.07.012PubMedGoogle ScholarCrossref
    19.
    Burtelow  M, Riley  E, Druzin  M, Fontaine  M, Viele  M, Goodnough  LT.  How we treat: management of life-threatening primary postpartum hemorrhage with a standardized massive transfusion protocol.  Transfusion. 2007;47(9):1564-1572. doi:10.1111/j.1537-2995.2007.01404.xPubMedGoogle ScholarCrossref
    20.
    Collins  PW, Cannings-John  R, Bruynseels  D,  et al; OBS2 study collaborators.  Viscoelastometry guided fresh frozen plasma infusion for postpartum haemorrhage: OBS2, an observational study.  Br J Anaesth. 2017;119(3):422-434. doi:10.1093/bja/aex245PubMedGoogle ScholarCrossref
    21.
    Netherlands Trial Register.  Transfusion strategies in women during major obstetric haemorrhage. https://www.trialregister.nl/trial/3909. Accessed July 19, 2019.
    22.
    Sentilhes  L, Vayssière  C, Deneux-Tharaux  C,  et al.  Postpartum hemorrhage: guidelines for clinical practice from the French College of Gynaecologists and Obstetricians (CNGOF): in collaboration with the French Society of Anesthesiology and Intensive Care (SFAR).  Eur J Obstet Gynecol Reprod Biol. 2016;198:12-21. doi:10.1016/j.ejogrb.2015.12.012PubMedGoogle ScholarCrossref
    23.
    Henriquez  DDCA, Bloemenkamp  KWM, Loeff  RM,  et al; TeMpOH-1 study group.  Fluid resuscitation during persistent postpartum haemorrhage and maternal outcome: a nationwide cohort study.  Eur J Obstet Gynecol Reprod Biol. 2019;235:49-56. doi:10.1016/j.ejogrb.2019.01.027PubMedGoogle ScholarCrossref
    24.
    Shakur  H, Roberts  I, Bautista  R,  et al; CRASH-2 trial collaborators.  Effects of tranexamic acid on death, vascular occlusive events, and blood transfusion in trauma patients with significant haemorrhage (CRASH-2): a randomised, placebo-controlled trial.  Lancet. 2010;376(9734):23-32. doi:10.1016/S0140-6736(10)60835-5PubMedGoogle ScholarCrossref
    25.
    WOMAN Trial Collaborators.  Effect of early tranexamic acid administration on mortality, hysterectomy, and other morbidities in women with post-partum haemorrhage (WOMAN): an international, randomised, double-blind, placebo-controlled trial.  Lancet. 2017;389(10084):2105-2116. doi:10.1016/S0140-6736(17)30638-4PubMedGoogle ScholarCrossref
    26.
    Andersen  LW, Granfeldt  A, Callaway  CW,  et al; American Heart Association’s Get With The Guidelines–Resuscitation Investigators.  Association between tracheal intubation during adult in-hospital cardiac arrest and survival.  JAMA. 2017;317(5):494-506. doi:10.1001/jama.2016.20165PubMedGoogle ScholarCrossref
    27.
    Andersen  LW, Raymond  TT, Berg  RA,  et al; American Heart Association’s Get With The Guidelines–Resuscitation Investigators.  Association between tracheal intubation during pediatric in-hospital cardiac arrest and survival.  JAMA. 2016;316(17):1786-1797. doi:10.1001/jama.2016.14486PubMedGoogle ScholarCrossref
    28.
    Austin  PC.  An introduction to propensity score methods for reducing the effects of confounding in observational studies.  Multivariate Behav Res. 2011;46(3):399-424. doi:10.1080/00273171.2011.568786PubMedGoogle ScholarCrossref
    29.
    Lu  B.  Propensity score matching with time-dependent covariates.  Biometrics. 2005;61(3):721-728. doi:10.1111/j.1541-0420.2005.00356.xPubMedGoogle ScholarCrossref
    30.
    Nakahara  S, Tomio  J, Takahashi  H,  et al.  Evaluation of pre-hospital administration of adrenaline (epinephrine) by emergency medical services for patients with out of hospital cardiac arrest in Japan: controlled propensity matched retrospective cohort study.  BMJ. 2013;347:f6829. doi:10.1136/bmj.f6829PubMedGoogle ScholarCrossref
    31.
    Xie  Y, Bowe  B, Li  T, Xian  H, Yan  Y, Al-Aly  Z.  Risk of death among users of proton pump inhibitors: a longitudinal observational cohort study of United States veterans.  BMJ Open. 2017;7(6):e015735. doi:10.1136/bmjopen-2016-015735PubMedGoogle Scholar
    32.
    Bolliger  D, Görlinger  K, Tanaka  KA.  Pathophysiology and treatment of coagulopathy in massive hemorrhage and hemodilution.  Anesthesiology. 2010;113(5):1205-1219. doi:10.1097/ALN.0b013e3181f22b5aPubMedGoogle ScholarCrossref
    33.
    Hancock  A, Weeks  AD, Lavender  DT.  Is accurate and reliable blood loss estimation the ‘crucial step’ in early detection of postpartum haemorrhage: an integrative review of the literature.  BMC Pregnancy Childbirth. 2015;15:230. doi:10.1186/s12884-015-0653-6PubMedGoogle ScholarCrossref
    34.
    Pallasmaa  N, Ekblad  U, Gissler  M, Alanen  A.  The impact of maternal obesity, age, pre-eclampsia and insulin dependent diabetes on severe maternal morbidity by mode of delivery: a register-based cohort study.  Arch Gynecol Obstet. 2015;291(2):311-318. doi:10.1007/s00404-014-3352-zPubMedGoogle ScholarCrossref
    35.
    Schols  SE, Lancé  MD, Feijge  MA,  et al.  Impaired thrombin generation and fibrin clot formation in patients with dilutional coagulopathy during major surgery.  Thromb Haemost. 2010;103(2):318-328. doi:10.1160/TH09-06-0396PubMedGoogle ScholarCrossref
    36.
    Henriquez  DDCA, Bloemenkamp  KWM, van der Bom  JG.  Management of postpartum hemorrhage: how to improve maternal outcomes [published online June 8, 2018].  J Thromb Haemost. doi:10.1111/jth.14200Google Scholar
    37.
    Austin  PC.  Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies.  Pharm Stat. 2011;10(2):150-161. doi:10.1002/pst.433PubMedGoogle ScholarCrossref
    38.
    Austin  PC.  A comparison of 12 algorithms for matching on the propensity score.  Stat Med. 2014;33(6):1057-1069. doi:10.1002/sim.6004PubMedGoogle ScholarCrossref
    39.
    Lunt  M.  Selecting an appropriate caliper can be essential for achieving good balance with propensity score matching.  Am J Epidemiol. 2014;179(2):226-235. doi:10.1093/aje/kwt212PubMedGoogle ScholarCrossref
    40.
    Rozé  JC, Cambonie  G, Marchand-Martin  L,  et al; Hemodynamic EPIPAGE 2 Study Group.  Association between early screening for patent ductus arteriosus and in-hospital mortality among extremely preterm infants.  JAMA. 2015;313(24):2441-2448. doi:10.1001/jama.2015.6734PubMedGoogle ScholarCrossref
    41.
    Janssen  KJ, Donders  AR, Harrell  FE  Jr,  et al.  Missing covariate data in medical research: to impute is better than to ignore.  J Clin Epidemiol. 2010;63(7):721-727. doi:10.1016/j.jclinepi.2009.12.008PubMedGoogle ScholarCrossref
    42.
    Moons  KG, Donders  RA, Stijnen  T, Harrell  FE  Jr.  Using the outcome for imputation of missing predictor values was preferred.  J Clin Epidemiol. 2006;59(10):1092-1101. doi:10.1016/j.jclinepi.2006.01.009PubMedGoogle ScholarCrossref
    43.
    Li  P, Stuart  EA, Allison  DB.  Multiple imputation: a flexible tool for handling missing data.  JAMA. 2015;314(18):1966-1967. doi:10.1001/jama.2015.15281PubMedGoogle ScholarCrossref
    44.
    Mitra  R, Reiter  JP.  A comparison of two methods of estimating propensity scores after multiple imputation.  Stat Methods Med Res. 2016;25(1):188-204. doi:10.1177/0962280212445945PubMedGoogle ScholarCrossref
    45.
    Penning de Vries  B, Groenwold  R.  Comments on propensity score matching following multiple imputation.  Stat Methods Med Res. 2016;25(6):3066-3068. doi:10.1177/0962280216674296PubMedGoogle ScholarCrossref
    46.
    Rubin  DB, Schenker  N.  Multiple imputation in health-care databases: an overview and some applications.  Stat Med. 1991;10(4):585-598. doi:10.1002/sim.4780100410PubMedGoogle ScholarCrossref
    47.
    Austin  PC.  Balance diagnostics for comparing the distribution of baseline covariates between treatment groups in propensity-score matched samples.  Stat Med. 2009;28(25):3083-3107. doi:10.1002/sim.3697PubMedGoogle ScholarCrossref
    48.
    Belitser  SV, Martens  EP, Pestman  WR, Groenwold  RH, de Boer  A, Klungel  OH.  Measuring balance and model selection in propensity score methods.  Pharmacoepidemiol Drug Saf. 2011;20(11):1115-1129. doi:10.1002/pds.2188PubMedGoogle ScholarCrossref
    49.
    Ali  MS, Groenwold  RH, Belitser  SV,  et al.  Reporting of covariate selection and balance assessment in propensity score analysis is suboptimal: a systematic review.  J Clin Epidemiol. 2015;68(2):112-121. doi:10.1016/j.jclinepi.2014.08.011PubMedGoogle ScholarCrossref
    50.
    Groenwold  RH, de Vries  F, de Boer  A,  et al.  Balance measures for propensity score methods: a clinical example on beta-agonist use and the risk of myocardial infarction.  Pharmacoepidemiol Drug Saf. 2011;20(11):1130-1137. doi:10.1002/pds.2251PubMedGoogle ScholarCrossref
    51.
    Normand  ST, Landrum  MB, Guadagnoli  E,  et al.  Validating recommendations for coronary angiography following acute myocardial infarction in the elderly: a matched analysis using propensity scores.  J Clin Epidemiol. 2001;54(4):387-398. doi:10.1016/S0895-4356(00)00321-8PubMedGoogle ScholarCrossref
    52.
    Haukoos  JS, Lewis  RJ.  The propensity score.  JAMA. 2015;314(15):1637-1638. doi:10.1001/jama.2015.13480PubMedGoogle ScholarCrossref
    53.
    Collins  PW, Cannings-John  R, Bruynseels  D,  et al.  Viscoelastometric-guided early fibrinogen concentrate replacement during postpartum haemorrhage: OBS2, a double-blind randomized controlled trial.  Br J Anaesth. 2017;119(3):411-421. doi:10.1093/bja/aex181PubMedGoogle ScholarCrossref
    54.
    Wikkelsø  AJ, Edwards  HM, Afshari  A,  et al; FIB-PPH trial group.  Pre-emptive treatment with fibrinogen concentrate for postpartum haemorrhage: randomized controlled trial.  Br J Anaesth. 2015;114(4):623-633. doi:10.1093/bja/aeu444PubMedGoogle ScholarCrossref
    55.
    Gillissen  A, van den Akker  T, Caram-Deelder  C,  et al; TeMpOH-1 Study Group.  Coagulation parameters during the course of severe postpartum hemorrhage: a nationwide retrospective cohort study.  Blood Adv. 2018;2(19):2433-2442. doi:10.1182/bloodadvances.2018022632PubMedGoogle ScholarCrossref
    ×