Low Molecular Weight Heparin Thromboprophylaxis or No Treatment following Cesarean Delivery: A Pilot Randomized Controlled Trial

Abstract

Objective

This study aimed to evaluate the feasibility of randomizing patients to weight-based low molecular weight heparin (LMWH) versus no pharmacologic thromboprophylaxis following cesarean delivery (CD).

Study Design

Single-center, open-label pilot randomized controlled trial of individuals aged 18+ undergoing CD at the University of Utah Health from November 2023 to June 2024. Those with a contraindication to anticoagulation, a plan for therapeutic anticoagulation, or considered at highest risk for postpartum venous thromboembolism (VTE; i.e., undergoing cesarean-hysterectomy, high-risk thrombophilia, personal history of thromboembolism) were excluded. Enrolled individuals were randomized in a 1:1 ratio utilizing block randomization with randomly varying block sizes to receive weight-based LMWH for 14 days or no pharmacologic thromboprophylaxis. The primary outcome was feasibility, defined as ≥35% enrollment of eligible individuals and retention of ≥85% of enrolled individuals through all study procedures. Secondary feasibility outcomes included the number of eligible patients per month, approach rate, enrollment rate, and retention rate. Additional outcomes included VTE, wound hematoma, patient-reported symptoms, or a bleeding complication within 6 weeks postpartum. Baseline characteristics were compared between those approached and enrolled and those not enrolled. The proportion meeting each of the outcomes was reported with 95% confidence intervals (CI).

Results

Over the 6-month study period, 694 patients were screened and found eligible for an average of 106 eligible patients per month. There were 611 patients approached (88%, 95% CI: 85.6–90.5), of which 64 enrolled (10.5%, 95% CI: 8–12.9), and 61 participants were retained through all study procedures (95.3%, 95% CI: 90–100). Thus, the overall primary outcome feasibility parameters were not met. Among the 64 individuals enrolled and randomized, the mean age was 31.0 years (standard deviation: 5.5 years), and the majority were non-Hispanic White (56%). Baseline characteristics were similar between those who were approached and enrolled compared with those not enrolled. There were no differences in additional clinical outcomes (VTE, wound hematoma, patient-reported symptoms, or bleeding complications) by prophylaxis group.

Conclusion

In this pilot trial, individual patient randomization to weight-based LMWH or no pharmacologic thromboprophylaxis after CD was not feasible due to low enrollment rates. Future trials addressing postpartum thromboembolism prevention should consider alternative study designs.

Introduction

Obstetric venous thromboembolism (VTE) contributes to nearly one in 10 maternal deaths in the United States, with risk peaking in the first 2 weeks postpartum and particularly accentuated after cesarean delivery (CD).^1–6 Pharmacologic prophylaxis with low molecular weight heparin (LMWH) is effective in reducing perioperative VTE in nonobstetric surgical fields.^7,8 Extrapolating to obstetrics, national guidelines recommend risk-based stratification of individuals after delivery to guide utilization of pharmacologic prophylaxis for VTE prevention.^2,9–11 However, the efficacy of LMWH to prevent VTE after CD remains unproven and may be associated with more bleeding complications.^12–15 Therefore, there is justifiable controversy, and practice is highly variable across the United States.¹⁶ Depending on the algorithm utilized at the delivery hospital of an individual, from less than one percent to over one-third of patients will receive pharmacologic prophylaxis.¹⁴ The need for a large-scale, rigorous clinical trial addressing the use of LMWH thromboprophylaxis to prevent VTE following CD has been identified.

Before a clinical trial utilizing individual patient randomization to address the efficacy and safety of LMWH after CD can be completed, the feasibility of such a trial needs to be evaluated. Randomization at the individual patient level will require over 60,000 participants to detect a clinically meaningful difference in rates of VTE by receipt of LMWH prophylaxis. Prior pilot trials outside of the United States have suggested such a trial design is not feasible.^17–20 However, the feasibility of individual patient randomization to LMWH prophylaxis or no pharmacologic prophylaxis in the United States during a modern period has not been tested.

Therefore, we aimed to evaluate the feasibility and acceptability of individual patient randomization to weight-based LMWH prophylaxis or no pharmacologic prophylaxis following CD at a single academic institution.

Materials and Methods

This was a single-center, open-label, parallel-group pilot randomized controlled trial (RCT) of weight-based LMWH prophylaxis compared with no pharmacologic prophylaxis after CD. Individuals aged 18 years or older and undergoing CD at the University of Utah Health from November 2023 to June 2024 were considered eligible. Individuals with a contraindication to anticoagulation, a plan for therapeutic anticoagulation, known renal dysfunction (defined as creatinine clearance < 30 mL/minute), receipt of antepartum anticoagulation for > 2 weeks, undergoing cesarean-hysterectomy, or with a personal history of VTE or high-risk thrombophilia were excluded. The inclusion and exclusion criteria were selected to reflect the anticipated population of a future trial evaluating the efficacy and safety of LMWH prophylaxis in low to moderate-risk groups. In such a future trial, those at the highest risk for a VTE (i.e., with a personal history of VTE, high-risk thrombophilia, or undergoing hysterectomy) are anticipated to be excluded with a recommendation for pharmacologic prophylaxis based on existing data.^2,3 Similarly, weight-based LMWH was selected as the intervention, as prior studies found superiority of weight-based compared to fixed dose LMWH to achieve prophylactic anti-Xa levels.^21–24

Eligible individuals were approached prior to delivery or within 12 hours postpartum (i.e., prior to the anticipated first dose of LMWH prophylaxis). Informed, written consent was obtained from all participants. Enrolled individuals were randomized to receive either weight-based LMWH or no pharmacologic therapy in a 1:1 ratio using a block randomization schema, in which block sizes varied randomly. The randomization schema was generated in SAS and then uploaded and maintained in REDCap for randomization upon enrollment. Participants, research staff, and clinicians were unmasked to study group allocation in this pragmatic design.

Participants randomized to weight-based LMWH prophylaxis received enoxaparin dosing at 0.5 mg/kg (rounded to the nearest 10 mg) based on admission weight, administered subcutaneously every 12 hours.²¹ Therapy was initiated 12 to 24 hours postdelivery in alignment with the American College of Obstetricians and Gynecologists and the Society for Obstetric Anesthesia and Perinatology guidelines.^2,25 Participants received therapy inpatient and at hospital discharge to complete a total course of 14 days. Participants randomized to no pharmacologic therapy received no LMWH prophylaxis, nor an alternative anticoagulant.

The remainder of intrapartum and postpartum care was per standard institutional practice. For those randomized to enoxaparin therapy, medication cost was billed through insurance as per the standard approach at our institution. For those with a remaining co-pay after insurance, or utilizing self-pay, medication costs were covered fully. Deferral of an enoxaparin dose or discontinuation of therapy early was at the discretion of the clinical care team. All patients received mechanical prophylaxis with sequential compression devices intraoperatively and postoperatively until ambulatory per institutional protocol across both study groups.

The primary outcome was feasibility, defined as 35% enrollment of eligible individuals and retention of 85% of enrolled individuals through all study procedures. The thresholds to define feasibility were informed by existing feasibility literature, a prior pharmacologic prophylaxis trial at the University of Utah Health, and prior international pilot studies on heparin-based thromboprophylaxis use.^{17–21,26,27} Secondary feasibility outcomes included the number of eligible patients per month, approach rate (proportion of eligible patients approached), enrollment rate (proportion of eligible patients enrolled), and retention rate (proportion enrolled completing all study procedures). The reasons for the nonapproach of eligible patients by the research staff were collected (e.g., no research staff available). Reasons for patient decline of enrollment and withdrawal from the study were also collected.

All participants were contacted at 2 weeks postpartum via phone call and asked about completion of the study procedures. For those in the enoxaparin group, this was verbal confirmation of LMWH prophylaxis adherence, and for those in the no pharmacologic prophylaxis group, this was verbal confirmation of no receipt of enoxaparin therapy. Both groups were additionally asked about any bruising or bleeding, wound complications, VTE events, or visits to an outside facility for care.

By design, this pilot trial did not address the efficacy of LMWH prophylaxis. However, data were collected on additional secondary outcomes, including VTE (deep vein thrombosis diagnosed by lower extremity Duplex ultrasound, pulmonary embolism diagnosed by either computed tomography angiography or ventilation-perfusion scan), wound hematoma (clinically diagnosed), or a bleeding complication (including readmission or return to the operating room) within 6 weeks postpartum.

Detailed medical record abstraction was completed by trained perinatal research staff for demographics, medical and obstetric history, delivery outcomes, and postpartum course. Final follow-up and outcomes were ascertained at 6 weeks postpartum. Adverse events (AEs) were monitored in real-time and reported to the investigator team and Institutional Review Board (IRB) as per protocol.

For all outcomes, the proportion meeting the outcome with a 95% confidence interval (CI) was reported. A descriptive summary of reasons for nonapproach by the research team, patient decline to participate, or participant withdrawal was also reported. Demographics were compared between patients approached and enrolled, and those approached but declined enrollment. Baseline characteristics and secondary clinical outcomes were compared between the randomization groups.

A power analysis was not performed as this pilot aimed to evaluate the feasibility and acceptability of identification, enrollment, randomization, and retention of the target population for a future trial rather than to test a hypothesis on efficacy.^28,29 The pilot trial was preset to enroll for 6 months.

The trial was registered at ClinicalTrials.gov (identifier: NCT06118957) and approved by the IRB of the University of Utah (approval no.: 00167480) in advance of study enrollment. The collected study data were managed using REDCap (Research Electronic Data Capture). All analyses were performed using SAS Version 9.4 (Cary, North Carolina, United States). Graphics were created using GraphPad Prism Version 9.4.1 (La Jolla, California, United States). The study was reported following the Consolidated Standards of Reporting Trials (CONSORT) guidelines (Supplementary Material 1, available in the online version).³⁰

Results

From November 17, 2023, to June 7, 2024, there were 694 eligible patients identified for an average of 106 patients per month. There were 611 patients (88.0%, 95% CI: 85.6–90.5) approached for enrollment. The reasons identified for not approaching the remaining 83 eligible patients included (not mutually exclusive) research staff not available to approach within 12-hour postpartum eligibility window (n = 47), clinician recommended against approach (n = 33), patient declined any research approach (n = 3), enrollment into another study without option for co-study enrollment (n = 2), or other (n = 1).

Of the 611 patients approached, 64 enrolled and were randomized (10.5%, 95% CI: 8–12.9; Fig. 1). The indicated reasons for patient decline to enroll included decline of any research (n = 197), not interested in this specific research (n = 56), enrolled in another study and declined participation in multiple studies (n = 10), did not want to potentially receive enoxaparin (n = 127), did not want to potentially not receive enoxaparin (n = 118), desired standard of care approach or to follow clinician recommendation (n = 9), reported feeling overwhelmed (n = 29), concerns about cost (n = 14), or other reason (n = 38). No enrolled participants withdrew from the trial.

Fig. 1.

CONSORT flow diagram.

The primary outcome feasibility threshold, defined by ≥35% enrollment of eligible individuals and retention of ≥85% of enrolled participants, was not met with an enrollment and retention rate of 10.5 and 95.3%, respectively (Fig. 2). Additional secondary feasibility outcomes are reported in Table 1. Baseline demographics did not differ between those approached who enrolled and those approached who declined enrollment (Table 2).

Fig. 2.

Rate of approach and enrollment of all screened eligible patients over the 6-month study period by month.

Table 1.

Primary and secondary feasibility outcomes

Outcome	Eligible (n = 694)	Range or % (CI)
No. of eligible patients per month	106	96.5–116.8
Approach rate (proportion of eligible patients approached)	n = 611/n = 694	88.0 (85.6–90.5)
Enrollment rate (proportion of eligible patients enrolled)	n = 64/n = 611	10.5 (8.0–12.9)
Retention rate (proportion enrolled completing study)	n = 61/n = 64	95.3 (90.0–100.0)

Abbreviation: CI, confidence interval.

Table 2.

Baseline demographics among those approached and enrolled compared with those declining enrollment

Characteristic	Enrolled; n = 64	Declined; n = 547	p-Value
Maternal age (y)	30.95 ± 5.5	31.37 ± 5.7	0.573
Maternal race
American Indian or Alaska Native	2 (3.1)	5 (0.9)	0.116
Asian	2 (3.1)	31 (5.7)	0.395
Black or African American	2 (3.1)	15 (2.7)	0.86
Native Hawaiian or other Pacific Islander	1 (1.6)	12 (2.2)	0.741
Caucasian	36 (56.3)	265 (48.4)	0.237
Other	14 (21.9)	155 (28.3)	0.274
Unknown/not documented	10 (15.6)	79 (14.4)	0.8
Maternal ethnicity			0.53
Hispanic or Latina	14 (21.9)	156 (28.5)
Non-Hispanic or Latina	40 (62.5)	315 (57.6)
Unknown/not documented	10 (15.6)	76 (13.9)

Note: Data are mean standard deviation (SD) or frequency (%). p-Value from a two-sample t-test for continuous or chi-square for categorical variable comparisons.

Of the 64 enrolled participants, 32 were randomized to LMWH prophylaxis, and 32 were randomized to no pharmacologic prophylaxis. Baseline characteristics were similar between the study groups (Table 3). Of those randomized to LMWH prophylaxis, 13 participants received medication co-pay cost support beyond medication coverage provided by insurance (range: $5–15). All trial participants received their allocated treatment during inpatient hospitalization. Of those enrolled, 61 were able to be contacted at approximately 2 weeks postpartum (95.3%, 95% CI: 90–100) and confirmed continued adherence to the assigned group outpatient. Patient-reported complications, including bleeding gums, nose bleeds, easy bruising, or abnormal vaginal bleeding, did not differ by randomization group. There were no reported VTE events in the full study population. Rates of wound hematoma and bleeding complications did not differ by randomization group (Table 4).

Table 3.

Baseline characteristics by randomized group

Characteristic	LMWH prophylaxis (n = 32)	No pharmacologic prophylaxis (n = 32)
Maternal age (y)	30.06 ± 6.1	31.84 ± 4.8
Maternal race
American Indian or Alaska Native	1 (3.1)	1 (3.1)
Asian	2 (6.3)	0 (0.0)
Black or African American	0 (0.0)	2 (6.3)
Native Hawaiian or other Pacific Islander	0 (0.0)	1 (3.1)
White	17 (53.1)	19 (59.4)
Other	11 (34.4)	3 (9.4)
Unknown	2 (6.3)	8 (25.0)
Maternal ethnicity
Hispanic or Latina	10 (31.3)	4 (12.5)
Not Hispanic or Latina	20 (62.5)	20 (62.5)
Unknown/not documented	2 (6.3)	8 (25.0)
Body mass index (kg/m2)	33.35 ± 7.4	35.83 ± 7.6
Co-morbid conditions
Chronic hypertension	3 (9.4)	2 (6.3)
Pregestational diabetes mellitus	0 (0.0)	2 (6.3)
Asthma	9 (28.1)	2 (6.3)
Autoimmune disease	2 (6.3)	1 (3.1)
Gestational diabetes mellitus	3 (9.4)	1 (3.1)
Hypertensive disorder of pregnancy	10 (31.3)	6 (18.8)
Fetal growth restriction	2 (6.3)	3 (9.4)
Tobacco use	1 (3.1)	5 (15.6)
Neonatal birth weight (grams)	3,189.06 ± 549.7	3,255.78 ± 601.1
Cesarean delivery in labor	19 (59.4)	12 (37.5)

Abbreviation: LMWH, low molecular weight heparin.

Note: Data are n (%) or mean ± standard deviation (SD),

Table 4.

Complications by study group

Outcome	LMWH prophylaxis (n = 32)	No pharmacologic prophylaxis (n = 32)	p-Value
Patient reported
Bleeding gums	1 (3.1)	2 (6.3)	0.554
Nose bleeds	1 (3.1)	1 (3.1)	0.99
Easy bruising	3 (9.4)	3 (9.4)	0.99
Abnormal vaginal bleeding	1 (3.1)	0 (0.0)	0.313
Venous thromboembolism	0 (0)	0 (0)	–
Wound hematoma	2 (6.3)	1 (3.4)	0.613
Bleeding complication	0 (0.0)	1 (3.4)	0.29

Abbreviation: LMWH, low molecular weight heparin.

Note: Data are n (%). p-Value not reported for venous thromboembolism as zero events were observed. Patient-reported complications were assessed by patients without clinical (in-office) assessment. Venous thromboembolism was defined by confirmed imaging, wound hematoma was clinically diagnosed, and bleeding complications included any readmission or return to the operating room within 6 weeks postpartum for a bleeding event.

There were seven AEs identifiedamongseven study participants. None were attributed as directly related to trial participation (Supplemental Appendix 1, available in the online version only).

Discussion

In this single-center pilot RCT, individual patient enrollment and randomization to weight-based LMWH or no pharmacologic thromboprophylaxis after CD were not feasible. The approach rate, enrollment rate, and retention rate were 88, 11, and 95%, respectively. There were no baseline demographic differences identified between those enrolling compared with those not enrolling. There were no VTE events identified during the study period. Neither wound nor bleeding complications differed by exposure to LMWH prophylaxis.

Whether administration of postpartum LMWH thromboprophylaxis prevents VTE events remains an unanswered question for clinicians. Similarly, whether such thromboprophylaxis increases the risk of bleeding complications (e.g., wound hematomas) is uncertain.^12–15 In a meta-analysis of eight randomized trials in general surgery comparing LMWH with placebo, or no treatment, LMWH prophylaxis resulted in a reduction in clinical VTE (RR: 0.29, 95% CI: 0.11–0.73).⁷ Despite robust data on the efficacy of LMWH to prevent VTE among general surgery and other sub-specialty surgical patients, similar trial data are not available for obstetric populations. Reflecting the uncertainties of available literature and varied expert opinions, current guidelines on postpartum LMWH prophylaxis defer decision-making to institutions, and experts have identified the need for rigorous prospective data.^2,9–11,16

Similarly, patients have also identified this need. Results from 122 participants delivering at hospitals in France and Switzerland, completing direct-choice exercises on potential benefits (reduction in VTE) and risks (bleeding complications) of pharmacologic thromboprophylaxis, found a high willingness to receive postpartum LMWH prophylaxis. Participants identified a desire to receive LMWH prophylaxis at a threshold of VTE risk as low as 0.1 to 0.2% even with a 1% risk for major bleeding from therapy.³¹ However, a large-scale trial has not been undertaken to date, in part, because of uncertain feasibility.

Prior pilot trials evaluating the feasibility of individual patient randomization to LMWH thromboprophylaxis versus placebo, or no treatment, after CD enrolled in primarily international clinical settings (i.e., Canada, Australia, and the United Kingdom) more than 10 years ago (i.e., enrollment across studies 1998–2013).^17–20 Burrows et al concluded that there was feasibility to upscale and complete a larger trial of LMWH prophylaxis based on enrollment of 76 patients at a single tertiary care center in Australia with a recruitment rate of 26%.¹⁷ The other three trials concluded non-feasibility based on low recruitment rates (6.6–11%).^18–20 More recently, Blondon et al evaluated the feasibility of recruitment to an open-label trial of postpartum LMWH for 10 days, or no therapy, in those considered at risk for VTE at the Geneva University Hospitals in Switzerland. Across 25 weeks of recruitment (May–November 2022), 77 patients were enrolled and randomized (n = 77/323, enrollment rate 23.8%, 95% CI: 19.2–38.9) with 84.6% retention and adherence to therapy.³² The authors concluded the potential feasibility of a multi-center trial in European countries. Our findings extend prior work to a modern period in the U.S. health system. Using predefined feasibility metrics, we concluded that it would not be feasible to complete a trial of LWMH efficacy with individual patient randomization.

The identified high approach rate (88%) reflects the ability of the research staff to screen and approach eligible patients efficiently. The University of Utah Obstetrics and Gynecology Research Network (OGRN) provides infrastructure and staffing for nearly 24-hour coverage of research enrollment, which likely contributed to these findings. The high rate of retention suggests participant engagement and willingness to complete study procedures, as well as reported adherence to treatment groups. The primary concern for upscaling the current trial methodology is centered around the low enrollment rate of eligible patients, which would limit the ability to achieve an adequate sample size for the outcome of a VTE event.

At least 60,000 participants would be needed for a future trial utilizing individual patient randomization to achieve statistical power to detect a clinically meaningful difference in rates of postpartum VTE by pharmacologic prophylaxis. This is assuming a base rate of VTE of 0.2% with a potential 35 to 70% risk reduction from LMWH prophylaxis use, as found in other surgical fields.^7,8 If 20 hospitals, each with approximately 5,000 deliveries annually enrolled in a multi-center trial with enrollment rates of 10 to 11%, it would take at least 6 years to achieve the necessary sample size. Thus, the use of individual patient randomization to LMWH thromboprophylaxis or no therapy is not practical to allow for completion of a multi-center rigorous trial with accrual of the necessary sample size in the United States.

A common reason for nonparticipation among eligible individuals was a strong desire to receive, or not receive, LMWH prophylaxis. This reflects strong preheld opinions on the receipt of LMWH prophylaxis. The rate of heparin-based thromboprophylaxis across 17 U.S. hospitals was previously found to be 12%, but this ranged widely by institution from as low as 0.21% to as high as 34.8%.¹⁴ We may suggest the current findings on variation in patient perception of necessity or nonnecessity of LMWH thromboprophylaxis may reflect the wider uncertainty, among patients, clinicians, and experts alike, on the efficacy and safety of this therapy.^16,33

This study has several strengths. The selected patient population, intervention, and methods for outcome ascertainment reflect an evidence-informed protocol that would be employed in a future, large-scale trial, if found feasible.^6,9,21 This study followed best practice in pilot trial development by evaluating feasibility and acceptability, which were necessary to inform future work in this area.^26,27

This study also has limitations. While our experience reflects feasibility at a single academic institution, our findings may not be generalizable to all hospital settings. In particular, the preexisting practice around use or nonuse (i.e., what is the local standard of care) may alter clinician and patient perception of the necessity of therapy and thereby alter patient enrollment. This was an open-label trial, and the comparator was no therapy rather than placebo (saline) injections, which may introduce bias. Participant verbal adherence was evaluated at the 2-week postpartum phone call, but bloodwork or other measures of adherence were not collected. However, such a pragmatic design would be necessary for a future trial.

While we determined a trial designed with the use of individual patient randomization to LMWH thromboprophylaxis or no therapy after CD is not feasible to upscale for a future, large multicenter study, prevention of postpartum VTE remains an imperative topic to be addressed. Our findings suggest that further evaluation of risks and benefits of LMWH thromboprophylaxis—a necessary undertaking—will require an alternative study design.

Supplementary Material

Supplementary Material is available at https://doi.org/10.1055/a-2753-9286.

Key Points.

• Individual patient randomization to enoxaparin or no therapy after CD was not feasible.

• The approach rate, enrollment rate, and retention rate were 88, 11, and 95%, respectively, in this single-center pilot.

• Future prospective studies may need to consider alternative designs.