Delayed-type hypersensitivity to low-molecular-weight heparin and cross-reactivity to fondaparinux and danaparoid in a twin pregnancy

Background

Venous thromboembolism (VTE) remains a leading cause of maternal morbidity and mortality. ¹ Pregnancy increases the risk of VTE four- to five-fold due to the physiological changes of gestation—namely venous stasis, vascular injury, and a hypercoagulable state. ¹ The absolute incidence of pregnancy-associated VTE is estimated at approximately 1–2 cases per 1000 pregnancies, with the highest risk in the first six weeks postpartum. ² Risk is further elevated in the presence of advanced maternal age, obesity, caesarean delivery, multiple gestation, thrombophilia, or prolonged immobility. ²

Low-molecular-weight heparins (LMWHs) are recommended by international guidelines as first-line agents for both prophylaxis and treatment of VTE during pregnancy, due to their excellent safety profile, minimal placental transfer, and predictable pharmacokinetics.^3–6 These features contribute to their preference over unfractionated heparin (UFH) and vitamin K antagonists such as warfarin. Delayed-type hypersensitivity (DTH) reactions with heparin have been reported in 7.5% of the general population. ⁷ The risk of heparin reaction is significantly higher in pregnancy. A prospective cohort study of 111 pregnant women reported that 19.8% receiving heparin developed allergic DTH reactions, with a median onset of 50 days after initial exposure (range 5–184 days). ⁸ These reactions typically present as pruritic, erythematous, or eczematous plaques at injection sites, and in some cases may progress to generalised cutaneous involvement.

This case highlights the complex diagnostic and therapeutic considerations encountered when heparin intolerance arises during pregnancy, particularly in the setting of limited treatment alternatives.

Case presentation

A 23-year-old woman in her first pregnancy with dichorionic diamniotic twins presented to an emergency department with chest pain and acute shortness of breath at 25 weeks of gestation. She had a body mass index of 37 kg/m² and no personal or family history of thrombosis or thrombophilia. She had a previous history of angioedema following exposure to shellfish. She described the pain as non-pleuritic and only lasting a few seconds before self-resolving intermittently over a 24-h period. Laboratory investigations revealed normal high sensitivity cardiac troponin-I and brain natriuretic peptide. The women declined pulmonary imaging because of concerns about fetal radiation exposure. Bilateral lower limb Doppler ultrasound was completed to evaluate for a DVT and revealed a 10 cm segment of thrombus within the proximal posterior tibial vein in the left leg. The woman was commenced on therapeutic enoxaparin (1 mg/kg twice daily).

After 21 days of treatment, she developed localised urticarial reactions at injection sites (Figure 1). Enoxaparin was ceased, and a trial of fondaparinux was conducted as an inpatient. Thirty minutes following administration of fondaparinux, the woman developed a widespread urticarial rash extending over the torso, back and upper and lower limbs (Figure 2). A trial of danaparoid was conducted, with worsening of the rash occurring 30 min post-exposure, consistent with cross-reactivity among heparin products. The woman was transferred to a tertiary referral centre for multidisciplinary review by haematology, dermatology, immunology and obstetric medicine services.

OPEN IN VIEWER

Figure 1.

Initial urticarial reaction following enoxaparin.

OPEN IN VIEWER

Figure 2.

Urticarial reaction following exposure to fondapariux and danaparoid.

Peripheral blood eosinophil count was mildly elevated with a peak of 0.68 × 10⁹/L (normal 0.04–0.5). C-reactive protein was 28 mg/L (<6). Skin biopsy demonstrated mild spongiotic dermatitis and moderate superficial perivascular dermatitis with scattered eosinophils and red-cell extravasation. Direct immunofluorescence was negative for autoimmune skin disease, consistent with the clinical diagnosis of a drug eruption. Intercellular cement substance and skin basement membrane tissue auto-antibody testing were negative. Total serum IgE, serum tryptase, platelet count and immunoglobulins were normal. Testing for antinuclear, antinuclear cytoplasmic and antithyroid antibodies was negative. Repeat Doppler ultrasound showed resolution of the thrombus. Echocardiography revealed normal right ventricular systolic pressure and absence of right ventricular systolic strain. Protein C, protein S and antithrombin III levels were in the normal range, accounting for physiological changes during pregnancy, and testing for antiphospholipid antibodies, factor V Leiden and prothrombin gene mutation were negative.

The patient's cutaneous symptoms were managed with loratadine 10 mg daily, a brief tapering course of prednisone (25 mg for five days, 10 mg for two days, then 5 mg for two days), topical hydrocortisone 0.5 % cream for facial lesions, topical triamcinolone 0.02 % for the trunk and limbs, and promethazine as required. Options for ongoing management of her DVT were discussed with the patient, including warfarin treatment, insertion of an inferior vena cava filter, intravenous heparin, enoxaparin with ongoing glucocorticoid therapy, or to cease all anticoagulation therapy and monitor with fortnightly lower limb Doppler ultrasound and symptomatic review. Repeat lower limb duplex ultrasound was completed 21 days after the original diagnosis, to evaluate the thrombus and guide counselling regarding ongoing VTE management. The ultrasound showed complete resolution of the DVT, and the woman opted for monitoring, deferring further therapeutic intervention. This included fortnightly lower limb Doppler ultrasounds and symptomatic review. There was no evidence of recurrent DVT on a fortnightly Doppler ultrasound. The woman delivered healthy infants of birthweights 2551 and 1582 g at 36 weeks’ gestation by caesarean section, performed because of breech presentation and fetal growth restriction. The woman was subsequently treated with oral rivaroxaban 10 mg daily for six weeks postpartum for VTE prophylaxis. She continued to breastfeed during this time. She has been referred to an allergy clinic for consideration of desensitisation to heparin.

Discussion

The case presented raises the diagnostic problem of skin reactions in pregnancy in the setting of LMWHs therapy, and the management issue of VTE in the setting of intolerance of heparin and heparinoids.

Heparin therapy may be associated with DTH reactions, immediate hypersensitivity reactions, bullous haemorrhagic dermatosis or skin necrosis. ⁷ DTH (type IV) reactions to unfractionated or LMWH typically present as localised erythema at injection sites, which may then progress to generalised eczematous or maculopapular eruption in 3–10% of cases. Risk factors in the general population include female sex, obesity and treatment duration greater than nine days. Immediate hypersensitivity reactions to heparin are rare, and manifest as IgE-mediated anaphylactic or non-IgE-mediated anaphylactoid reactions, presenting with localised or generalised urticaria, angioedema, bronchospasm, hypotension and tachycardia. Skin necrosis may occur with heparin-induced thrombocytopenia, characterised by the appearance of erythaematous patches 3–15 days after initiation of therapy that progress to skin necrosis. ⁹ Bullous haemorrhagic dermatosis is an uncommon reaction to LMWH characterised by the appearance of rapidly expanding haemorrhagic bullae after days to weeks of heparin therapy.

The onset of pruritic, erythematous plaques at enoxaparin injection sites was consistent with a DTH reaction. The timing of onset, absence of elevated serum IgE and initial localisation of reaction to injection sites were not consistent with atopic eruption of pregnancy. Similarly, the distribution of lesions was inconsistent with pruritic urticarial papules and plaques of pregnancy. The absence of vesicobullous lesions, negative immunofluorescence and negative tissue autoantibodies were not consistent with pemphigoid gestationis, pemphigoid, pemphigus, epidermis bullosa or bullous lupus erythematosus.

In individuals with DTH reactions to LMWH, cross-reactivity with other LMWHs based on intradermal testing occurs in 50–80% of cases, and with UFH in 24%. ¹⁰ Where DTH reactions occur with heparin in pregnancy, a trial of fondaparinux is indicated.^11–13 Cross-reactivity or concomitant sensitisation to fondaparinux was seen in only two of 48 cases (4.2%) following LMWH-DTH reactions. ¹² A systematic review and meta-analysis including more than 550 pregnancies with exposure to fondaparinux estimated the incidence rate of rash and allergic reactions to be 1.74%, and the incidence rate of injection site reactions to be 1.85%. ¹³ Compared with LMWH, fondaparinux did not increase the incidence of vaginal bleeding and postpartum haemorrhage and was associated with a reduced risk of hepatic transaminase elevation, gastrointestinal symptoms, allergies and injection site skin reactions. ¹³

An analysis of clinical outcomes in 91 case reports of pregnancies in 83 women where danaparoid was used for heparin intolerance (30 with current heparin-induced thrombocytopenia; 17 with past heparin-induced thrombocytopenia; 44 with other heparin intolerances) reported 49 maternal serious adverse events in 42 pregnancies, including two maternal deaths, four non-fatal haemorrhages and six new thromboembolic episodes. ¹⁴ Of 25 LMWH/heparin skin reactions, 15 resolved with a change to danaparoid therapy, while 10 persisted with danaparoid therapy requiring cessation. A prospective study comparing the efficacy of danaparoid and UFH (each combined with aspirin) in women with obstetric antiphospholipid syndrome reported no cases of allergic reactions in 22 women treated with danaparoid, compared with two allergic reactions in 42 women (4.8%) treated with UFH. ¹⁵

Intravenous challenge tests with UFH revealed the absence of reaction in all of 85 individuals with a history of DTH reactions to LMWH.^12,16–18

Saraiva et al. described a woman whose pregnancy was complicated by DVT, who developed DTH reaction to enoxaparin and tinzaparin at injection sites. ¹⁹ The woman was subsequently managed with prednisolone 20 mg/day and dalteparin. An attempt to reduce the prednisolone dose at 30 weeks’ gestation resulted in new skin lesions. The woman was maintained on prednisolone 20 mg/day and dalteparin until delivery at 37 weeks’ gestation, then for six weeks postpartum.

There is a lack of safety data regarding the use of direct oral anticoagulants (DOACs) in pregnancy. A retrospective cohort study identified 614 unique cases of DOAC exposure during pregnancy between 2007 and 2020. ²⁰ The median duration of DOAC exposure was 5.3 weeks into pregnancy. Of the 336 pregnancies for which pregnancy outcomes were available, 21 (6%) showed fetal abnormalities, of which 12 (4%) were adjudicated as major birth defects potentially related to DOAC exposure. Estimated relative infant doses of rivaroxaban in the setting of maternal rivaroxaban therapy while breastfeeding have ranged between 0.82 and 5%, with infant plasma rivaroxaban plasma concentrations below the lower limit of quantification, suggesting breastfeeding is acceptable with maternal rivaroxaban therapy.^21–24

Warfarin is a teratogen with the greatest risk being to fetuses exposed between six- and 12-weeks’ gestation. Characteristic features of warfarin teratogenicity involve abnormal development of bone and cartilage, with nasal and bone hypoplasia, and stippled epiphyses. ²⁵ Other manifestations of warfarin embryopathy include frontal bossing, short neck, short limbs, polydactyly, low birth weight and respiratory compromise due to choanal atresia. ²⁶ Second and third trimester exposure to warfarin (fetopathy) has been linked to a variety of neurological abnormalities, including optic atrophy, mental retardation, intellectual impairment, microcephaly, cerebral agenesis, hypotonia, spasticity, Dandy-Walker malformation, deafness and seizures. ²⁶ A total of 20 cases of neurologic manifestations secondary to intrauterine warfarin exposure were described in a review in 2007. ²⁶ Potential causes of neurological sequelae following intra-uterine exposure to warfarin include fetal intracerebral haemorrhage, a direct toxic effect of warfarin, or neurologic sequelae secondary to skeletal abnormality. Information regarding the risk of warfarin embryopathy/fetopathy has been derived from literature concerning anticoagulation in the setting of mechanical heart valves. The risk of teratogenicity/fetopathy appears to be dose dependent. Risk of fetal embryopathy/fetopathy with maternal warfarin dose less than 5 mg/day has been estimated to be 2.3%, compared with a risk of 12.4% with maternal warfarin dose greater than 5 mg/day. ²⁷ A systematic review of English- and Chinese-language databases disclosed 12 cases of fetal intracranial haemorrhage with maternal warfarin therapy. Only five live births occurred, and two of the infants had normal neurological development at 12 months of age. Maternal warfarin dose was less than 5 mg/day in four cases. ²⁸

A review of 199 cases of insertion of inferior vena cava filters (IVC-F) during pregnancy for acute VTE found at least one complication of insertion in 33 women (19%), including two maternal deaths. ²⁹ The authors concluded IVC-F should only be considered in pregnancy in the setting of acute proximal DVT shortly prior to delivery, with a contraindication for anticoagulant therapy, or progression of DVT despite anticoagulant therapy. ²⁹

Conclusion

This case highlights the diagnostic and therapeutic challenges posed by heparin allergy in pregnancy. It underscores the need for early multidisciplinary collaboration, particularly involving obstetric medicine, haematology, dermatology, and allergy/immunology. In the absence of consensus guidelines for anticoagulation in the context of confirmed heparin hypersensitivity, shared decision-making, informed by individual thrombotic risk, allergy confirmation and gestational age, remains critical.

Future directions include the development of safer, non-heparin-based anticoagulants with validated safety in pregnancy and breastfeeding. There is also a need for updated national guidelines addressing this rare but clinically significant complication.