Certolizumab pegol to prevent adverse pregnancy outcomes in patients with antiphospholipid syndrome and lupus anticoagulant (IMPACT): results of a prospective, single arm, open-label, phase 2 trial

Abstract

Objectives

The aim of this study was to evaluate whether certolizumab pegol, a TNF-α inhibitor with little or no transport across the placenta, added to standard treatment with low molecular weight heparin plus low dose aspirin, reduces rates of adverse pregnancy outcome (APO) in high-risk pregnancies with antiphospholipid syndrome (APS).

Methods

We assessed treatment with certolizumab in pregnant patients with APS and lupus anticoagulant, administered gestational weeks’ 8 through 28, in addition to standard treatment. The primary APO was a composite of fetal death ≥10 weeks’ gestation or preeclampsia with severe features or placental insufficiency requiring delivery <34 weeks’ gestation. Target sample size was 45 with expected APO rate of 20% with certolizumab versus 40% in historical controls from a prospectively-observed population of similarly managed APS pregnancies.

Results

Fifty-one patients were enrolled, and 9 had primary APO (17.6%; 95% CI: 8.4%−30.9%). Excluding 6 patients that had a pregnancy loss <10 weeks’ gestation or fetal loss due to genetic abnormalities, primary APO occurred in 9 of 45 patients (20%; 95% CI: 9.6%−34.6%), meeting pre-determined criteria for efficacy of certolizumab and significantly lower than rates in historical controls. Median gestational age at delivery in certolizumab-treated patients was 36.5 weeks and was after 30 weeks in those who met the primary outcome of preeclampsia. Neonatal survival to hospital discharge was 93%. There were no serious infections and no new cases or severe flares of lupus.

Conclusions

Certolizumab appears effective in preventing placenta-mediated adverse outcomes in high-risk APS patients.

Trial registration number:

NCT03152058.

INTRODUCTION

Antiphospholipid syndrome (APS) is an autoimmune thrombo-inflammatory disorder that occurs most commonly in women of reproductive age and is associated with adverse pregnancy outcomes (APO), including fetal death and preterm birth due to preeclampsia with severe features or placental insufficiency, and with arterial and venous thromboses.^{1, 2} The presence of a lupus anticoagulant (LA) is strongly associated with higher risk for clinical obstetric APS.^3–6 The recommended treatment of pregnant patients with APS is low molecular weight heparin (LMWH) plus low dose aspirin (LDA).^{7, 8} Despite treatment, APS patients positive for LA suffer a 39–86% rate of placenta-mediated APO.^3–6

APOs associated with APS are due to inadequate vascularization of the developing placenta resulting in malperfusion of the intervillous space by maternal blood and placental hypoxia leading to maternal manifestations of preeclampsia and fetal manifestations of placental insufficiency.⁹ An animal model that mimics APS pregnancy in humans revealed that poor placental development is due primarily to inflammation.^{10, 11} Antiphospholipid antibodies target placental tissue and activate complement, leading to the recruitment and activation of leukocytes, release of proinflammatory mediators, angiogenic imbalance, and ultimately abnormal placentation and fetal death. TNF-α is a critical downstream effector of placental dysfunction and fetal death in preclinical models.¹² Importantly, TNF-α blockade normalized placentation, prevented inflammation at the maternal-fetal interface, and rescued pregnancies in a murine model of APS-associated pregnancy complications.¹² In a mouse model of non-autoimmune, spontaneous preeclampsia, inhibition of TNF-α also prevented fetal deaths and growth restriction and normalized spiral artery remodeling and placental size and structure, avoiding the typical features of preeclampsia.¹³

The use of TNF-α antagonists has been reported to induce autoantibodies in 10–20% of patients.^14–17 Transient increases in anti-double stranded DNA (anti-dsDNA) antibodies are rarely associated with clinical manifestations of SLE.¹⁴ Induction of SLE-like disease is probably a class effect, but it is rare, mild, transient, and resolves after discontinuing treatment.^15–18 Increases in antiphospholipid antibodies have been described, and there are rare case reports of associated thrombotic events.^{17, 19}

To evaluate the capacity of TNF-α blockade, added to a regimen of LMWH plus LDA, to decrease placenta-mediated APO in APS patients, we studied certolizumab pegol, an Fc-free, pegylated Fab’ fragment of a humanized TNF blocking monoclonal antibody that has minimal or no transfer across the placenta and it is safe and well tolerated in human pregnancy.^20–22 Here we report the results of the IMPACT (Improve Pregnancy in APS with Certolizumab Therapy) trial, an open label phase 2 trial of certolizumab pegol to prevent preeclampsia with severe features and placental insufficiency in pregnant patients with APS with LA. This is the first trial of a biologic therapy to prevent maternal complications of pregnancy that can lead to fetal death or preterm delivery.

METHODS

Study design

The IMPACT trial was a single-arm open-label, phase 2 trial of certolizumab pegol to prevent preeclampsia with severe features and placental insufficiency in pregnant patients with APS with LA recruited from clinical practices in the United States and Canada. The protocol is provided in the Supplementary Material. The use of certolizumab in the study was approved by the U.S. Food and Drug Administration (Investigational New Drug #130908). Ethics approval in the United States was provided by the University of Utah Institutional Review Board (#00094818) and in Canada by Mount Sinai Hospital Research Ethics Board (#17–0219-A). An independent data and safety monitoring board was appointed by the NIH to review the trial every 6 months. This trial is registered with ClinicalTrials.gov: NCT03152058.

The trial was designed by the co-investigators and funded by the National Institute of Arthritis and Musculoskeletal and Skin Diseases, Lupus Foundation of America, with administrative support and medication provided by UCB Pharma Inc., Brussels, Belgium. Drs. Branch and Salmon had full access to the trial results and supervised preparation of the manuscript. The trial statistician developed the study design, statistical analysis plan and performed the analyses. The funders played no role in the conduct of the trial, data analyses, or drafting the manuscript. The authors vouch for the completeness and accuracy of the data and for the fidelity of the trial to the protocol.

Patients

Eligible patients were 18–40 (+364 days) years of age, pregnant <8 weeks’ gestation and had the following characteristics: APS by revised Sapporo criteria;¹ positive for LA, on two or more occasions 12 or more weeks apart, one of which must be within the previous 18 months; and hematocrit >26%. Patients were excluded if they had hypertension at enrollment (BP >140/90 mmHg), multifetal gestation, pre-existing diabetes, autoimmune conditions requiring maintenance prednisone >10 mg daily, platelet count <100,000 cells per microliter, serum creatinine >1.2 mg/dL, proteinuria (>500 mg per day or spot urine protein-to-creatinine ratio ≥0.5), and contraindications to certolizumab (protocol).²³ Patients gave written informed consent.

All patients were treated with LMWH and LDA under the supervision of their physician. In general, those with a history of thromboembolism received therapeutic dose LMWH; those without such a history received thromboprophylactic doses. Some patients also were prescribed hydroxychloroquine at their physician’s discretion.

Procedures

Certolizumab was administered as a dose of 400 mg subcutaneously beginning by 8 weeks and 6 days of gestation, 400 mg 2 and 4 weeks later, and 200 mg every other week until 28 weeks’ gestation. The first dose of certolizumab was administered between 7 weeks 0 days (after initial obstetric ultrasound confirmation of a singleton viable pregnancy) and 8 weeks 6 days, because we were targeting abnormal placentation which begins early in the pregnancy. The PROMISSE Study demonstrated that that angiogenic dysregulation, an indication of placental dysfunction, was evident as early as 12–15 weeks’ gestation.²⁴ Treatment was discontinued before 28 weeks’ gestation to limit medication exposure when additional benefits were unlikely. Study personnel contacted patients regarding certolizumab injections every 2–3 weeks and recorded dates of medication administration.

All patients enrolled were under the care of or had consultation by a maternal-fetal medicine subspecialist. Screening occurred at <8 weeks’ gestation, and maternal care visits were at least monthly, with serial fetal sonographic assessments after mid-second trimester. Fetal surveillance testing was done at least weekly ≥32 weeks’ gestation. Patients had complete blood counts and anti-dsDNA testing every 4–6 weeks. Other assessments are in the protocol. The COVID-19 pandemic disrupted this schedule in 7 patients. Study personnel contacted patients at least monthly regarding clinical status and signs or symptoms of systemic lupus erythematosus (SLE). Pregnancy outcomes were ascertained from physicians and patients and verified by medical records.

Outcomes

The primary outcome was a composite of placenta-mediated APO: (1) otherwise unexplained fetal death ≥10 weeks’ gestation, or (2) preeclampsia with severe features or placental insufficiency requiring delivery prior to 34 weeks’ gestation. Preeclampsia with severe features was diagnosed according to American College of Obstetricians and Gynecologists guidelines (ACOG). Placental insufficiency was defined according to the revised Sapporo criteria for APS as abnormal or non-reassuring fetal surveillance test(s), defined as (a) non-reactive non-stress test or biophysical profile score <6, suggestive of fetal hypoxemia, (b) abnormal Doppler flow velocimetry waveform analysis suggestive of fetal hypoxemia, e.g. absent or reverse end-diastolic flow in the umbilical artery, (c) oligohydramnios, e.g. an amniotic fluid index of <5 cm or single deepest vertical pocket less than 2 cm, or (d) ultrasound estimated fetal weight ≤5th percentile.¹ Fetal deaths were reviewed with regard to causation by Dr. Robert M. Silver.²⁵

Secondary clinical outcomes included preeclampsia or placental insufficiency not requiring delivery <34 weeks’ gestation, preterm labor or preterm rupture of membranes resulting in delivery prior to 36 weeks’ gestation, gestational age at delivery (completed weeks’ gestation at birth), small-for-gestational age birthweight (<10th percentile), neonatal death due to complications of prematurity, maternal thrombosis, and known adverse reactions to certolizumab. Patients were monitored for serious adverse events and adverse events as defined in the protocol.

Statistical analysis

This trial was designed as a single-stage phase 2 trial with 90% power at a 1-sided Type I error rate of 5% to reject the null hypothesis that the APO rate with treatment is greater than or equal to 40% under the alternative hypothesis that the true APO rate with treatment is 20%. The assumed APO rate of 40% under the null hypothesis was considered to be conservative by the study team given the observed APO rate of 44% among IMPACT-eligible patients (i.e., with APS and LA) in PROMISSE (N=61). The PROMISSE study was a prospective multicenter observational study of pregnancies in patients with antiphospholipid antibodies or APS and/or systemic lupus erythematosus (SLE) that enrolled 700 patients between 2003–2013 in North America (protocol).^3,26

The target sample size for this trial was 45 participants, and the treatment would be considered efficacious if 12 or fewer participants out of 45 have an APO. To account for a 10% drop-out rate, the goal was to enroll a total of 50 participants. The APO rate and corresponding exact 95% confidence intervals were computed in the following three analysis populations: (1) intent-to-treat (ITT) or full analysis population: all enrolled participants regardless of compliance to study medication; (2) modified ITT (mITT): participants who did not have a pregnancy loss before 10 weeks’ gestation or due to chromosomal abnormalities; (3) per-protocol (PP): participants who strictly adhered to the protocol. In exploratory analyses, gestational age, APO, and live birth rates in the most recent prior pregnancies of IMPACT patients that were maintained beyond 10 weeks gestational age and treated with LMWH and LDA were also compared to those in study pregnancies using the Wilcoxon signed rank test and exact McNemar’s test, respectively.

Role of the funding source

The trial was designed by the co-investigators who had had full access to the trial results and supervised preparation of the manuscript. The trial statistician developed the study design, statistical analysis plan and performed the analyses. The funders played no role in the conduct of the trial, data analyses, or drafting the manuscript.

Patient and public involvement

Patients and/or public were not involved in the design, conduct, reporting or dissemination of this research.

RESULTS

From May 2017 to February 2024, seventy-six patients were assessed for eligibility, and 51 patients were enrolled (Figure 1). Of the 51 ITT pregnancies, 2 had pregnancy losses <10 weeks’ gestation and 4 ended in fetal death attributable to genetic causes, leaving 45 patients in the mITT population. One patient was non-compliant with treatment, leaving 44 in the PP population.

FIGURE 1:

IMPACT Enrollment Flow Diagram

Baseline demographic and clinical characteristics of the 51 enrolled patients are shown in Table 1. Among 51 ITT patients, 26 (51%) and 23 (49%) respectively, were moderate-to-high positive^{1, 2} for IgG anticardiolipin or anti-beta2-glycoprotein-1 antibodies, and 17 (34%) were positive for both. Fifteen patients were LA positive only. Twenty-nine (57%) met thrombotic and obstetric criteria for APS;¹ 45 (88%) met obstetric criteria and 35 (69%) met thrombotic criteria. Ten (20%) patients had SLE. ITT, mITT and PP populations did not differ with regard to demographic and clinical characteristics.

Table 1:

Demographic and Clinical Characteristics of IMPACT Patients

	Intention-to-Treat (N=51)	Modified Intention-to-Treat (N=45)	Per-Protocol (N=44)
Age, years, median (range)	32 (20–40)	32 (20–40)	32 (20–40)
Body mass index, kg/m2, median (range)	25.4 (18.7–44.9)	25.7 (18.7–44.9)	25.5 (18.7–44.9)
<25, N (%)	25 (49)	21 (47)	21 (48)
25–30, N (%)	14 (27)	13 (29)	13 (30)
>30, N (%)	12 (24)	11 (24)	10 (23)
Ethnicity
Hispanic/Latina, N (%)	10 (20)	10 (22)	10 (23)
Race
White, N (%)	40 (78)	35 (78)	35 (80)
Asian, N (%)	4 (8)	3 (7)	2 (5)
Black, N (%)	1 (2)	1 (2)	1 (2)
>1 Race, N (%)	2 (4)	2 (4)	2 (5)
Other/Not Reported, N (%)	4 (8)	4 (9)	4 (9)
Lupus anticoagulant (LA), N (%)	51 (100)	45 (100)	44 (100)
Anticardiolipin (aCL) and LA, N (%)	28 (55)	24 (53)	23 (52)
IgG positive*	26 (51)	22 (49)	21 (48)
IgM only positive*	2 (4)	2 (4)	2 (5)
Anti-beta-2 glycoprotein-1† (aB2GP1) and LA, N (%)	25 (51)	21 (47)	20 (45)
IgG positive*	23 (49)	19 (42)	18 (41)
IgM only positive*	2 (4)	2 (4)	2 (5)
aCL or aB2GP1† and LA, N (%)	35 (70)	30 (67)	29 (66)
IgG positive	32 (64)	27 (60)	25 (57)
IgM only positive	3 (6)	3 (7)	3 (7)
aCL and aB2GP1† and LA positive, N (%)	18 (37)	15 (33)	15 (34)
IgG positive	17 (34)	14 (31)	13 (30)
IgM only positive	1 (2)	1 (2)	1 (2)
LA positive, aCL and aB2GP1 negative, N (%)	15 (30)†	15 (33)	15 (34)
Systemic lupus erythematosus, N (%)	10 (20)	8 (18)	8 (18)
With renal involvement, N (%)	1 (2)	1 (2)	1 (2)
History of thrombotic features of APS, N (%)	35‡ (69)	32‡ (71)	31 (70)
Venous thromboembolism, N (%)	30 (59)	27 (60)	26 (59)
Arterial thromboembolism, N (%)	6 (12)	6 (13)	5 (11)
Stroke, N (%)	4 (8)	4 (9)	3 (7)
Other,§ N (%)	2 (4)	2 (4)	2 (5)
History of obstetric features of APS, N (%)	45 (88)	39 (87)	38 (86)
History of both thrombotic and obstetric features of APS, N (%)	29 (57)	26 (58)	25 (57)
History of immune thrombocytopenia, N (%)	4 (8)	4 (9)	4 (9)
History of hemolytic anemia, N (%)	2 (4)	2 (4)	2 (5)
History of thyroid disease, N (%)	8 (16)	7 (16)	6 (14)
Medications at enrollment
Low molecular weight heparin, N (%)	50 (98)ǁ	45 (100)	44 (100)
Thromboprophylactic, N (%)	21 (41)	20 (44)	20 (45)
Therapeutic, N (%)	29 (57)	25 (56)	24 (55)
Low dose aspirin, N (%)	50 (98)ǁ	45 (100)	44 (100)
Hydroxychloroquine, N (%)	31 (61)	26 (58)	26 (59)
Azathioprine, N (%)	2 (4)	2 (4)	2 (5)
Prednisone,¶ N (%)	3 (6)	1 (2)	1 (2)
Other,** N (%)	2 (4)	2 (4)	2 (5)

^*Defined as > 40 standard units in most recent assay result

^†aB2GP1 results unavailable for one Intention-to-Treat patient

^‡One patient had a history of both arterial and venous thrombosis

^§Includes 2 patients with amaurosis fugax.

^ǁOne patient miscarried (<10 weeks’ gestation) prior to starting low molecular weight heparin and aspirin

^¶Prednisone <10 mg daily.

^**Includes 1 patient on tacrolimus and 1 patient on labetalol and nifedipine.

At enrollment, treatments included LDA in all patients and thromboprophylactic or therapeutic dose LMWH in 21 (41%) and 29 (57%) patients, respectively (Table 1). Thirty-one (61%) patients were also on hydroxychloroquine.

Among the 49 ITT patients who had prior pregnancies, 43 (88%) had a history of a prior IMPACT APO (Table 2A). Collectively, these patients had 135 prior pregnancies, 87 of which were maintained beyond 10 weeks, and of those, 60 (69%) resulted in an IMPACT APO (Table 2B). Thirty-nine of these pregnancies treated with LMWH and LDA were maintained beyond 10 weeks, 27 (69%) of which resulted in an IMPACT APO, and 15 (38%) had live births with infants surviving to discharge.

Table 2.

Obstetric History of IMPACT Patients

A. Individual Patients (N=49)*
Gravidity excluding study pregnancy, median (range)	2 (1–10)
History of ≥3 consecutive early losses (<10 weeks’ gestation)†, N (%)	3 (6)
History of fetal death(s) ≥ 10 weeks’ gestation, N (%)	38 (78)
History of preeclampsia or placental insufficiency requiring delivery <34 weeks’ gestation, N (%)	36 (73)
History of a prior IMPACT adverse pregnancy outcomef‡, N (%)	43 (88)
History of live birth surviving to hospital discharge, N (%)	18 (37)
B. Prior Pregnancy Outcomes
	All prior pregnancies	LDA and LMWH§ treated prior pregnancies	LDA, LMWH§ and HCQ treated prior pregnancies
Total Pregnancies, N	135ǁ	56¶	16**
Pregnancies beyond 10 weeks’ gestation, N	87	39	12
Fetal death ≥ 10 weeks gestation, N (% of pregnancies beyond 10 weeks’ gestation)	53 (61)	21 (54)	8 (67)
Preeclampsia or placental insufficiency requiring delivery <34 weeks’ gestation, N (% of pregnancies beyond 10 weeks’ gestation)	43 (49)	24 (62)	9 (75)
IMPACT adverse pregnancy outcome‡, N (% of pregnancies beyond 10 weeks’ gestation)	60 (69)	27 (69)	9 (75)
Live birth surviving to hospital discharge, N (% of pregnancies beyond 10 weeks’ gestation)	29 (33)	15 (38)	3 (25)

^*Two patients had no prior pregnancies

^†In 2 of the 3 patients, this was the only obstetric criterion for APS

^‡Otherwise unexplained fetal death and/or preeclampsia or placental insufficiency requiring delivery <34 weeks

^§Defined as low dose aspirin and heparin agent started in 1st trimester of pregnancy

^ǁTotal number of pregnancies among 49 intention-to-treat patients with prior pregnancies

^¶Total number of pregnancies among 26 patients treated with LDA and LMWH

^**Total number of pregnancies among 8 patients treated with LDA, LMWH and HCQ, 4 of whom had SLE

IMPACT primary APO occurred in 9 of 51 patients in the ITT population (17.6%; 95% CI: 8.4%−30.9%) and 9 of 45 patients in the mITT population (20%; 95% CI: 9.6%−34.6%) (Table 3), meeting the pre-determined criteria for efficacy of certolizumab (≤ 12 APO out of 45). Eight primary outcomes occurred in 44 PP pregnancies (18.2%; 95% CI: 8.2%−32.7%). Given that the upper bounds of the 95% CIs for the true APO rate are less than 35% in all analysis populations, the APO rates are significantly lower than the 40% APO rate assumed for the null hypothesis, 44% historical control APO rate observed in IMPACT-eligible PROMISSE patients and 37.0% rate among the PROMISSE patients who received both LMWH and LDA (Figure 2).

Table 3.

IMPACT Outcomes

	Intention-to-Treat	Modified Intention-to-Treat	Per-Protocol
Primary Outcome	N = 51	N = 45	N = 44
Otherwise unexplained fetal death ≥10 weeks’ gestation	3	3	2
Preeclampsia or placental insufficiency requiring delivery <34 weeks’ gestation	7	7	6
Composite primary outcome N (%)	9* (17.6)	9* (20.0)	8 (18.2)
Infant survival	N = 51	N = 45	N = 44
Live births	42 (82)	42 (93)	42 (95)
Live births surviving to hospital discharge	42 (82)	42 (93)	42 (95)
Non-obstetric Secondary Outcomes	N = 51	N = 45	N = 44
Maternal thrombosis N (%)	1 (2)	1 (2)	1 (2)
Known adverse reaction to certolizumab	0	0	0
Obstetric Secondary Outcomes	N = 42†	N = 42†	N = 42†
Gestational age at delivery‡
With primary outcome, median (range)	31 (27–33)	31 (27–33)	31 (27–33)
Without primary outcome, median (range)	37 (32–39)	37 (32–39)	37 (32–39)
Preterm live births (<37 weeks’ gestation)
With primary outcome, N (%)	6 (14)	6 (14)	6 (14)
Without primary outcome, N (%)	12 (29) §	12 (29) §	12 (29) §
Preeclampsia or placental insufficiency not requiring delivery <34 weeks’ gestation, N (%)	2 (5) ǁ	2 (5) ǁ	2 (5) ǁ
Small-for-for gestational age birthweight (<10th percentile)
With primary outcome, N (%)	3 (7)	3 (7)	3 (7)
Without primary outcome, N (%)	5 (12)	5 (12)	5 (12)
Neonatal death due to preterm birth for preeclampsia or placental insufficiency, N (%)	0	0	0

^*One participant met both outcomes: intrauterine fetal death at 21 weeks and preeclampsia with severe features

^†Obstetric secondary outcomes are only applicable to live births

^‡Completed weeks’ gestation at birth

^§Includes 1 case each of preterm labor at 32 weeks’ gestation, maternal pulmonary hypertension at 35 weeks’ gestation, flare of polychondritis at 35 weeks’ gestation, and suspected fetal growth restriction at 36 weeks’ gestation; 2 cases of fetal growth restriction (34 weeks’ and 36 weeks’ gestation, respectively), 2 cases of cholestasis of pregnancy at 36 weeks’ gestation; and 4 cases of patient-physician shared-decision elective preterm delivery at 35 or 36 weeks’ gestation.

^ǁBoth cases were diagnosed with fetal growth restriction and were delivered of small-for-gestational age infants; however, neither had preeclampsia

FIGURE 2:

IMPACT Trial Illustration

Standard of care = Low molecular weight heparin and low dose aspirin

APO = Adverse pregnancy outcome: otherwise unexplained fetal death >10 weeks gestation and/or severe Preeclampsia or placental insufficiency requiring delivery prior to 34 weeks gestation.

* Based on total number of pregnancies

Of 9 primary outcomes, 2 were fetal deaths at 10- and 16-weeks’ gestation. The 10-week loss was preceded by 2–3 weeks of vaginal bleeding. The 16-week loss was not accompanied by preeclampsia or placental insufficiency, genetic abnormality, or significant placental pathology. A third fetal death occurred at 22 weeks’ gestation, preceded by preeclampsia with severe features, in a patient who received only four doses of certolizumab and discontinued medication at 16 weeks. The six remaining primary outcomes were preeclampsia with severe features and/or placental insufficiency occurring at 27 to 33 weeks’ gestation; all delivered liveborn infants who survived to discharge. Excluding early pregnancy losses and fetal losses due to lethal genetic abnormalities, live births occurred in 42 of 45 mITT pregnancies (93%; 95% CI: 81.7%−98.6%), and all liveborn infants survived to hospital discharge (Table 3).

In exploratory analyses, we compared prior pregnancies in IMPACT patients. Thirty-nine prior pregnancies were treated with LMWH and LDA and progressed beyond 10 weeks’ gestation, such that an IMPACT primary outcome could have occurred (Table 2B). Of these, 27 (69%) resulted in an IMPACT APO and only 15 (38%) were liveborn and survived to hospital discharge. Comparing the most recent prior pregnancies treated with LMWH and LDA that progressed beyond 10 weeks with the patient’s certolizumab-treated pregnancy (Figure 3, N=24), IMPACT APO rates were 79% vs 21%, respectively (p<0.001), and the median gestational ages of delivery were 24.0 weeks versus 36.5 weeks, respectively (p<0.001). Furthermore, 8 of 24 (33%) prior pregnancies resulted in livebirths surviving to discharge compared to 20 of 24 (83%) in the subsequent certolizumab-treated pregnancies (p = 0.0002) (Figure 3).

Figure 3:

Paired outcomes, IMPACT pregnancy and most recent prior treated pregnancy beyond 10 weeks’ gestation

Pre-defined secondary outcomes are shown in Table 3. The only maternal thrombotic event was a minor ischemic stroke in a patient on therapeutic dose LMWH; the patient fully recovered (Table 4). There were no known adverse reactions to certolizumab. There were no cases of preeclampsia after 34 weeks’ gestation. Ultrasonographic features of placental insufficiency prompted preterm delivery in 2 patients after 34 weeks’ gestation (at 34- and 36-weeks’ gestation).

Table 4.

IMPACT Serious Adverse Events^*

Maternal Events†, n=8	Gest. Age at Time of Event (weeks)	Description
Arthralgias and hemolytic anemia	14	Patient with beta-thalassemia minor and known to be Coomb’s positive with minimal fall in hemoglobin from 10·1 to 9·6 g/dL. Hospitalized and treated with oral glucocorticoids.
Preterm labor	32	Spontaneous preterm labor
Pulmonary hypertension	34	At enrollment, the patient denied a history of cardiac disease. Her new diagnosis of pulmonary hypertension at 34 weeks’ gestation was attributed to multiple factors including SLE, prior pulmonary embolisms, morbid obesity, and obstructive sleep apnea. Labor induction at 35 weeks’ gestation undertaken to facilitate medical management. Fetal bradycardia in the second stage of labor led to urgent cesarean during which intraoperative blood loss of approximately 3 L resulted in maternal hypotension and severe maternal hypoxemia. Though appropriate intensive care unit measures were taken the patient succumbed to cardiopulmonary collapse several hours after delivery. The liveborn infant was of appropriate weight for the gestational age and is alive and well.
Shortness of breath, dry cough, low-grade fever	34	Upper respiratory infection with bronchospasm. Hospitalized and treated with antibiotics and steroid inhaler. CT angiogram negative for thrombosis.
SLE flare	31	Fatigue, arthralgias, transaminitis, and mild thrombocytopenia. Patient normotensive and without proteinuria. Hospitalized and treated with intravenous glucocorticoids and discharged on prednisone.
Thromboembolism	33	Minor ischemic stroke at 33 weeks’ gestation characterized by mild-to-moderate left upper and lower extremity sensory changes and imaging findings involving the posterior right insula and right parietal lobe. Mild thrombocytopenia and anemia were present at admission to hospital. Following adjustment of LMWH dose, symptoms dissipated, and the patient fully recovered.
Transaminitis/thrombocytopenia	34	Hospitalized for evaluation. Clinical features attributed to influenza A and an adverse drug reaction to azathioprine. Patient normotensive and without proteinuria.
Upper respiratory symptoms	33	Hospitalized for 48 hours. Negative tests for RSV, influenza, and SARS-Co-2. Treated with antibiotics.
Neonatal Events, n=2	Age of infant at time of Event (days)	Description
Meconium ileus and volvulus	3rd day of life	Preterm infant delivered at 32 weeks’ gestation secondary to preeclampsia. Surgical intervention requiring ileostomy. Readmitted 3 months after birth for ileal atresia and underwent enterectomy with reanastomosis and take-down of ileostomy.
RSV Infection	39th day of life	Preterm infant delivered at 31 weeks’ gestation secondary to preeclampsia. Readmitted with RSV and hospitalized for 20 days.

^*Serious Adverse Events are defined as any adverse event that:

• Results in death
• Is life threatening, or places the participant at immediate risk of death from the event as it occurred
• Requires or prolongs hospitalization
• Causes persistent or significant disability or incapacity
• Results in congenital anomalies or birth defects
• Is another condition which the investigators judge to represent significant hazards

^†Does not include hospitalizations (maternal and neonatal) that resulted in a primary outcome, unless the investigators judge otherwise.

There were no serious infections and no new diagnoses of SLE during the trial. TNF-α blockade has been associated with induction of anti-dsDNA and rarely clinical SLE. Eight of 51 had anti-dsDNA at enrollment, 4 of whom did not have a diagnosis of SLE and did not develop clinical SLE during the trial. Six patients converted from negative to low positive anti-dsDNA; none developed symptoms of SLE, and 3 reverted to negative. No patients with IgG anticardiolipin or anti-beta2-glycoprotein-1 antibodies had an increase in titer, and none converted from negative to positive. One patient increased from low-positive to high-positive IgM anticardiolipin without clinical consequences.

Serious adverse events and adverse events are shown in Tables 4 and 5, respectively. There was one death in a patient with SLE, prior pulmonary emboli, morbid obesity, and obstructive sleep apnea who was diagnosed with pulmonary hypertension at 34 weeks’ gestation and died due to cardio-pulmonary arrest after emergency cesarean at 35 weeks’ gestation. She had no evidence of thrombosis or SLE at autopsy. The healthy, liveborn infant was of appropriate weight. One patient had a minor ischemic stroke without sequalae. One patient had a SLE flare classified as severe and characterized by fatigue and arthralgias at 31 weeks’ gestation which responded to glucocorticoids. A second patient had a moderate flare at 8 weeks’ gestation that responded to glucocorticoids. No other patients developed symptoms of SLE-like disease or had flares. Serious adverse events occurred in 2 neonates, both of whom survived (Table 4).

Table 5.

IMPACT Adverse Events

Maternal Events	Mild*	Moderate†
Injection site reaction	4	1
Headache	4
URI	5
Anxiety	3
UTI	2
SLE Flare in pre-existing SLE		1
Vaginal yeast infection	1
COVID-19		1
Otherwise unexplained diarrhea	1
Abdominal pain: Constipation	1
Migraine aura/amaurosis-like episodes		1
Insomnia	1
Chest discomfort	1
Dry cough	1
Reaction to 2nd dose of Moderna COVID Vaccine (fever, body aches and nausea)		1
Polychondritis flare in pre-existing polychondritis		1
Ocular migraine	1
Rash	1
Cholestasis of pregnancy	4
Gestational Diabetes	2
Neonatal Events	Mild*	Moderate†
Eczema	1

^*Mild: An event that is easily tolerated by the patient, causing minimal discomfort and not interfering with everyday activities.

^†Moderate: An event that is sufficiently discomforting to interfere with normal everyday activities.

DISCUSSION

Certolizumab added to a regimen of LMWH and LDA appeared effective in preventing fetal death or delivery <34 weeks’ gestation due to preeclampsia with severe features or placental insufficiency in patients with APS and LA based on a priori assumptions that these events occur in 40% of such pregnancies. In IMPACT, the rate of this primary outcome was 20% in our mITT population, significantly lower than the rates in historical controls observed in PROMISSE patients eligible for IMPACT (44% in all pregnancies; 37% in those treated with LMWH and LDA). There were no safety signals related to certolizumab, such as serious maternal or neonatal infections or onset or flare of SLE.

The APO rate of 20% using certolizumab was much lower than that in prior pregnancies in our study patients (69% in all prior; 79% in most recent prior pregnancy). Moreover, APOs, when they occurred, were later in gestation in certolizumab-treated pregnancies compared to those in the patients’ prior treated pregnancies resulting in improved survival of neonates (93% vs 38%, respectively).

Despite the societal burden of fetal loss, preterm delivery, and maternal morbidity associated with such outcomes, treatment trials for APS pregnancy with LA are uncommon, small, and have not yielded new therapies. Retrospective studies of APS patients with LA treated with LMWH and LDA confirm an alarmingly high rate of second- or third-trimester placenta-mediated APO.^{5, 6, 27} Previously published prospective treatment trials of pregnancies in patients with antiphospholipid antibodies have included patients predominantly with recurrent early miscarriage, rather than preeclampsia and/or placental insufficiency, and many were either not tested or were negative for LA and/or did not meet international clinical or laboratory classification criteria for APS.^{28, 29} A 2020 Cochrane systematic review concluded that though treatment with a heparin agent and low dose aspirin may improve the live birth rate, characteristics of participants and adverse events were not uniformly reported, and the evidence was judged to be of low certainty.³⁰

There have been few trials with other therapeutic agents in APS pregnancy and none with definitive evidence of success. A randomized trial of intravenous immune globulin versus placebo in 16 patients with APS, 15 of them with LA, did not find a difference in obstetric outcomes, but the small size of the trial limited conclusions.³¹ Administration of pravastatin at the onset of preeclampsia and fetal growth restriction in 11 individuals with APS, who also received LMWH and LDA, improved outcomes compared to 10 who did not receive pravastatin, but subsequent studies of non-autoimmune patients did not show statistically significant clinical effects of pravastatin.³² We are aware of only one previous study of TNF-α blockade in APS pregnancy. This case series included 12 individuals with obstetric APS undergoing IVF because of recurrent miscarriage, six of whom were LA positive. Patients with arterial or venous thrombosis were excluded. TNF-α blockade during the first trimester was added to LMWH and LDA and yielded six live births.³³

A strength of our study is the prospective acquisition of patients meeting international classification criteria for APS¹ and carrying the highest risk phenotype for APO marked by repeatedly positive LA. We used a rare disease trial model with only one US recruiting site, and enrolled patients from 16 states and one Canadian province over seven years. In addition, we recruited extremely high-risk patients that were committed to the study; no one was lost to follow-up, despite being in a trial during the COVID pandemic.

Importantly there was no evidence of clinically significant changes in autoantibodies or serious exacerbations of autoimmune disease. One patient, with a history of pulmonary emboli and a subsequently negative echocardiogram, was discovered to have pulmonary hypertension 6 weeks after the last dose of certolizumab. The patient succumbed to severe hypoxemia secondary to hemorrhage resulting in hypovolemia during emergency cesarean in the second stage of labor, complications unrelated to certolizumab.

The trial also had limitations, including the absence of a randomized control group. We chose a single arm design because offering no treatment to women, many of whom had suffered multiple serious APOs, was unethical as their childbearing years were limited. In addition, APS with LA is a rare disease³⁴ challenging the feasibility of a larger randomized trial. To compensate, we considered comparable PROMISSE patients who would have been eligible for IMPACT as well as prior pregnancies in IMPACT patients to estimate control outcome rates in the absence of certolizumab. In addition, the majority of patients were White, a distribution similar to prevalence of APS in the US and Europe.³⁵ A third limitation is that 61% of our patients were receiving hydroxychloroquine at enrollment compared to 20–33% in previous pregnancies and in IMPACT-eligible PROMISSE patients. This shift in care likely results from the 2020 American College of Rheumatology conditional recommendation to use hydroxychloroquine in APS pregnancies which was based on limited or conflicting data.^{36, 37} Large registry studies suggest no benefit of hydroxychloroquine in SLE patients at risk for preeclampsia,^{38, 39} while a recent meta-anlysis suggested that hydroxychloroqine reduced preeclampsia and other APOs in SLE patients.⁴⁰ The contribution of APS to pregnancy outcomes was not considered. Moreover, one retrospective analysis that suggested benefit of hydroxychloroquine in a subset of high-risk patients with LA showed a 39% rate of APOs,⁶ significantly higher than the rate in our study. It is difficult to make a statement about potential efficacy of hydroxychloroquine in the IMPACT study population given the limited sample size and possibility of confounding by indication when comparing hydroxychloroquine exposed and non-exposed patients. Our study was not powered for subgroup analyses by hydroxychloroquine use.

In this single-arm study, the first phase 2 trial of a biologic to prevent placenta-mediated APO, the addition of certolizumab resulted in substantial and clinically meaningful reduction in APOs. These efficacy results, together with safety data and preclinical models,^{13, 41} support further studies with TNF-α blockade in conditions associated with high risk for placenta-mediated APOs. Our study heralds a new era for trials with biologics to prevent placenta-mediated adverse pregnancy outcomes, including preeclampsia, a life-threatening condition of global significance with limited effective treatments.

Supplementary Material

Protocol

KEY MESSAGES.

What is already known on this topic

Antiphospholipid syndrome (APS) is an autoimmune disorder associated with thrombosis and placenta-mediated adverse pregnancy outcomes (APO), including fetal death, preeclampsia, and fetal growth restriction, despite standard treatment with low molecular weight heparin (LMWH) plus low dose aspirin (LDA). Prior studies reported that the presence of lupus anticoagulant (LA) increases APO risk to rates greater than 40%.

In murine models of APS and non-autoimmune spontaneous preeclampsia, inhibition of tumor necrosis factor-alpha (TNF-α) prevented fetal deaths and growth restriction and normalized spiral artery remodeling and placental size and structure. Thus, TNF-α is a critical effector of placental dysfunction and fetal death in preclinical models of APS.

What this study adds

Certolizumab pegol is an Fc-free, pegylated Fab’ fragment of a humanized TNF-α blocking monoclonal antibody that has minimal or no transfer across the placenta and is safe and well tolerated in human pregnancy.

This is the first prospective treatment trial in pregnant patients using an immunomodulatory agent to prevent APOs. In this single arm, open-label, phase 2 trial, we assessed treatment with certolizumab administered during gestational weeks’ 8 through 28, in addition to LMWH and LDA, in pregnant patients with APS and LA and, thus, at high risk for adverse pregnancy outcomes. The primary APO was a composite of fetal death ≥10 weeks’ gestation or preeclampsia with severe features or placental insufficiency requiring delivery <34 weeks’ gestation.

The study met pre-determined criteria for efficacy of certolizumab and suggests that the addition of certolizumab pegol treatment to standard of care regimen can improve pregnancy outcomes in APS patients. In addition, we did not observe enhanced autoimmunity or disease flares.

How this study might affect research, practice or policy

The trial supports the concept that targeting inflammation, rather than thrombosis, is effective in preventing pregnancy complications in high-risk pregnant patients with APS. Clinicians should consider the addition of certolizumab pegol treatment beginning in the early first trimester in high-risk APS patients to reduce the rate of placenta-mediated adverse outcomes. Biomarkers to identify responders would be useful and might uncover other targets to protect pregnancies. Studies to determine whether TNF-α blockade prevents placenta-mediated adverse outcomes in other at-risk populations are warranted.

DATA SHARING

The NIH policy on data sharing can be found online: https://sharing.nih.gov/data-management-and-sharing-policy/about-data-management-and-sharing-policies/data-management-and-sharing-policy-overview#before