Worldwide, increasing numbers of postpartum women are receiving low-molecular-weight heparin for venous thromboembolism (VTE) prophylaxis. Five national guidelines recommend liberal low-molecular-weight heparin in a large proportion of birthing women, including most women delivering by cesarean.1 The rationale behind this is simple: pregnant and postpartum women are hypercoagulable, and deep vein thrombosis and pulmonary emboli cause morbidity and mortality. However, good intentions do not equal good evidence. Venous thromboemboli are rare and difficult to predict, and most episodes occur in women with one or no risk factors. Relative risks from observational studies have been translated into risk-assessment models, which select women to receive low-molecular-weight heparin.2 However, risk factors are common and VTE is rare. Women with common risk factors (eg, obesity, age older than 40 years, cesarean delivery) have an absolute VTE risk of 0.2% or less, and risk assessment modeling reveals a very low positive predictive value (0.13%).3 Only 1 in 800 high-risk women is destined to experience a VTE without prophylaxis, and the extent to which low-molecular-weight heparin might prevent that VTE is unknown. No risk-assessment model has been clinically validated, and there is no cohort or experimental evidence that low-molecular-weight heparin prophylaxis provides benefit for most women who qualify for it.
Guidelines from the American College of Obstetricians and Gynecologists (ACOG) and the American Society of Hematologists are notable exceptions to this trend. Noting poor evidence, they restrict recommendations for low-molecular-weight heparin to the few women with a history of VTE or potent thrombophilia, in whom the risk of VTE is 2–5% and the utility of low-molecular-weight heparin is widely acknowledged.4,5 The proportion of women eligible for low-molecular-weight heparin after cesarean delivery using American Society of Hematologists and ACOG guidelines is less than 1%, compared with 73% using guidelines from the United Kingdom.6
Without evidence or expert consensus, the study by Lu and colleagues7 in this month's issue of Obstetrics & Gynecology (see page 530) provides much-needed guidance. Benefit and harm from low-molecular-weight heparin were measured before and after introduction of a typical risk-assessment model in a cohort of 24,000 women, making this the first study to have enough statistical power to measure small differences in the incidence of VTE, bleeding, and wound complications. Women with a potent thrombophilia or history of prior VTE received heparin during both periods and were excluded.
The proportion of birthing women receiving low-molecular-weight heparin increased from 1.2% to 15.6%. The incidence of VTE did not change: 15 of 11,799 (0.13%) patients before and 16 of 12,430 (0.13%) patients after the risk-assessment model was introduced. However, the incidence of wound hematoma, unplanned surgery, and blood transfusion increased significantly, by 0.3%, 0.5%, and 0.9%, for numbers needed to harm of 333, 200, and 105, respectively; overall, an extra 240 women experienced harm after introduction of the protocol. Among 8,000 women delivering by cesarean, the number receiving low-molecular-weight heparin increased from 2.6% to 32.1%. The incidence of VTE was unchanged (0.16% vs. 0.25%), whereas the incidence of wound hematoma, unplanned surgery, and blood transfusion increased significantly by 0.9%, 1.0%, and 2.1%, respectively (numbers needed to harm 118, 102, and 49).
This is the best evidence available on the real-world performance of an obstetric risk-assessment model. Performed at a single center with a comprehensive electronic medical record, the quality of the data is high. There were 12 maternal deaths—none from pulmonary embolus. This is not surprising, because fewer than 1% of obstetric VTE are fatal pulmonary emboli; this supports previous estimates that the number needed to treat to prevent one fatal pulmonary embolism with postpartum low-molecular-weight heparin after cesarean delivery is 400,000–1,000,000.1,8 Given that 4% of women in this study experienced wound and bleeding complications from low-molecular-weight heparin after cesarean delivery, an estimated 16,000 to 40,000 women would be expected to be harmed for every pulmonary embolism death prevented. It would be presumptuous to assume that none of these complications would be fatal. The drug and nursing costs of low-molecular-weight heparin typically exceed $100 per person, yielding a primary cost of at least $40 million to prevent one fatal pulmonary embolism, without accounting for the cost of more than 10,000 additional reoperations, wound complications, and blood transfusions.
A drop in the incidence of fatal pulmonary embolism in the United Kingdom from 2003 to 2006 has been cited as evidence that liberal use of low-molecular-weight heparin saves lives; however, epidemiologic data at this scale are inadequate to suggest, let alone prove, causality. A modest reduction in overall mortality might be associated with improved prophylaxis for women with a prior VTE or history of thrombophilia, not necessarily those with common risk factors, or it might be due to normal statistical variation. By 2014, the incidence of fatal pulmonary embolism had increased to 15 per million, where it has remained since (Fig. 1)9—unchanged from 1982, 1985, 1988, 1991, and 2000, before the introduction of chemoprophylaxis. It is noteworthy that a dramatic fivefold drop in pulmonary embolism deaths from 1955 to 1982 (from 80 to 15 per million) accompanied the abandonment of bedrest—liberally prescribed to pregnant and postpartum women for decades by well-meaning physicians who lacked evidence.10
In the era of evidence-based medicine, it is remarkable that low-molecular-weight heparin prophylaxis has been widely disseminated without measurement of net benefit or harm. Cochrane reviewers have repeatedly called for large-scale randomized trials,11 and the American Society of Hematology concluded in 2018 that, “…the certainty of any benefit with provision of prophylaxis to antepartum or postpartum women with clinical risk factors is low or very low because of the lack of appropriate randomized studies and imprecision of the risk estimates.”5
Proponents of liberal chemoprophylaxis claim that randomized trials are not feasible. Because the benefit of low-molecular-weight heparin is so small, the required sample size is enormous.12 At the same time, they claim that the risk of VTE is so great that widespread administration of low-molecular-weight heparin is justified without evidence. This is a nonsensical argument: the risk cannot be simultaneously too small to measure and yet too great to make measurement unnecessary. Moreover, it is not necessary to prove mortality benefit with low-molecular-weight heparin if there is a reduction in nonfatal VTE, which occurs 100 times more often and would require a much smaller sample.
Proponents of low-molecular-weight heparin have also minimized harm from low-molecular-weight heparin, which Lu and colleagues have now demonstrated. Absent randomized trials, Lu and colleagues have done the next best thing: a large cohort study. The results are sobering. With a sample size of 24,000 patients, the power to detect a small reduction in VTE from low-molecular-weight heparin was limited; however, there was not a trend to suggest that benefit might exist. Yet harm from low-molecular-weight heparin was statistically and clinically significant. This lays the arguments of low-molecular-weight heparin enthusiasts to rest. Without further evidence, no one in good conscience can say to a birthing woman with common risk factors that she will benefit from low-molecular-weight heparin. Before chemoprophylaxis can be expanded beyond those few women identified in American Society of Hematologists and ACOG guidelines, net benefit needs to be demonstrated—ideally in randomized trials or, alternatively, in large cohort studies such as this one.
In light of the harm from low-molecular-weight heparin, low-dose acetylsalicylic acid (aspirin) may be an attractive alternative that is cheaper and easier to administer. Increasingly used after orthopedic surgery instead of more potent anticoagulants, acetylsalicylic acid may also be effective for birthing women. The international PARTUM (Postpartum Aspirin to Reduce Thromboembolism Undue Morbidity) randomized trial is underway to compare low-dose acetylsalicylic acid with placebo for postpartum women with moderate risk factors.13 Whether acetylsalicylic acid or low-molecular-weight heparin, evidence is needed that chemoprophylaxis is safe and effective before guidelines endorse it.
1. Kotaska A. Postpartum venous thromboembolism prophylaxis may cause more harm than benefit: a critical analysis of international guidelines through an evidence-based lens. BJOG 2018;125:1109–16. doi: 10.1111/1471-0528.15150
2. Jacobsen AF, Skjeldestad FE, Sandset PM. Incidence and risk patterns of venous thromboembolism in pregnancy and puerperium—a register-based case–control study. Am J Obstet Gynecol 2008;198:233.e1–7. doi: 10.1016/j.ajog.2007.08.041
3. Ban L, Sultan AA, West J, Tata LJ, Riley RD, Nelson-Piercy C, et al. External validation of a model to predict women most at risk of postpartum venous thromboembolism: maternity clot risk. Thromb Res 2021:S0049-3848(21)00360-1. doi:10.1016/j.thromres.2021.05.020
4. Thromboembolism in pregnancy. ACOG Practice Bulletin No 196. American College of Obstetricians and Gynecologists. Obstet Gynecol 2018;132:e1–17. doi: 10.1097/AOG.0000000000002706
5. Bates SM, Rajasekhar A, Middeldorp S, McLintock C, Rodger MA, James AH, et al. American Society of Hematology 2018 guidelines for management of venous thromboembolism: venous thromboembolism in the context of pregnancy. Blood Adv 2018;2:3317–59. doi:10.1182/bloodadvances.2018024802
6. Federspiel JJ, Wein LE, Addae-Konadu KL, Darwin KC, Talamo LE, et al. Venous thromboembolism incidence among patients recommended for pharmacologic thromboembolism prophylaxis after cesarean delivery in selected guidelines. J Thromb Haemost 2021;19:830–8. doi: 10.1111/jth.15218
7. Lu MY, Blanchard CT, Ausbeck EB, Oglesby KR, Page MR, Lazenby AJ, et al. Evaluation of a risk-stratified, heparin-based, obstetric thromboprophylaxis protocol. Obstet Gynecol 2021;138:530–8. doi: 10.1097/AOG.0000000000004521
8. Sibai BM, Rouse DJ. Pharmacologic thromboprophylaxis in obstetrics: broader use demands better data. Obstet Gynecol 2016;128:681–4. doi: 10.1097/AOG.0000000000001656
9. NPEU. Saving lives, improving mothers' care: lessons learned to inform maternity care from the UK and Ireland Confidential Enquiries into Maternal Deaths and Morbidity 2016-18. Accessed July 29, 2021. https://www.npeu.ox.ac.uk/mbrrace-uk/reports
10. Why mothers die 2000–2002. The sixth report of the Confidential Enquiries into Maternal Deaths in the United Kingdom. RCOG Press; 2004.
11. Middleton P, Shepherd E, Gomersall JC. Venous thromboembolism prophylaxis for women at risk during pregnancy and the early postnatal period. The Cochrane Database of Systematic Reviews 2021, Issue 3. Art. No.: CD001689. doi: 10.1002/14651858.CD001689.pub4
12. Friedman AM, D'Alton ME. Expert review: prevention of obstetric venous thromboembolism. Am J Obstet Gynecol 2021:S0002-9378(21)00548-2. doi: 10.1016/j.ajog.2021.05.004
13. U.S. National Library of Medicine.History of changes for study: NCT04153760. Pilot PARTUM trial: Postpartum Aspirin to Reduce Thromboembolism Undue Morbidity (PARTUM). Accessed July 29, 2021. https://clinicaltrials.gov/ct2/history/NCT04153760?V_2=View