Previous venous thromboembolism and thrombophilia

Women with previous thrombotic events have a higher risk of recurrence in pregnancy and postpartum. Although available studies are heterogeneous in design and enrolled a relatively small number of women, the risk can be stratified: the highest risk of recurrence is intuitively in women with previous recurrent venous thromboembolism (VTE), although the rate is unknown. Recurrence rates for women with a history of a single previous episode of thrombosis are 5.8–6.2% overall during pregnancy, equally distributed along trimesters and 8.3–10% in the postpartum period. The risk is higher for previous pregnancy-related or oestrogen-provoked events: 9.5–10%, compared with 2.7% if the previous event was not pregnancy- or oestrogen-related. Although the exact recurrence rates vary between studies, recent data support a higher risk attributable to this group. The smallest risk is associated with a previous event, which was provoked by a temporary event, no longer present.

Thrombophilia can be inherited (antithrombin, protein C, protein S deficiency, factor V Leiden and prothrombin gene variant) or acquired (antiphospholipid syndrome, including lupus anticoagulant or anticardiolipin antibodies). Twenty to 50 per cent of women with thrombosis have a thrombophilia. Again, the risk of VTE associated with each of these factors varies widely, and is also dependent on previous thrombotic history. The risk associated with asymptomatic defects is small, with the exception of those with antithrombin deficiency and combinations of defects.

Obesity

Obesity has emerged in recent years as an independent and important risk factor for TED. Although the risk associated with obesity is moderate, it is prevalent in women of childbearing age; 18% of those aged 25–34 years and 22% of those aged 35–44 years fell into this category in 2003 in the UK. The risk is most likely stratified, becoming stronger, the higher the body mass index. In the recent Confidential Enquiry of the 16 women who died from pulmonary embolism, three women had a BMI greater than 25, nine had a BMI greater than 30, including one with a BMI greater than 40. How obesity causes thrombosis is not entirely clear, and has been the subject of many studies. An association between obesity and increased hypercoagulability certainly exists as is reflected in increased thrombin generation. Evidence is emerging that levels of PAI-1 are significantly raised in obese individuals, with consequent inhibition of fibrinolysis. In addition, oxidative stress has been associated with adipose tissue. This leads to platelet activation, endothelial damage and shredding of activated platelet and endothelial cell derived microparticles, which in turn are thrombogenic.

Deep-vein thrombosis

Most DVT will occur in the lower limb and the pelvis. Significant differences exist in the location of the DVT in pregnant women compared with non-pregnant women. Eighty-two per cent of thrombi are left sided and, importantly, 71% of clots are in the proximal veins, with 64% in the iliac or femoral veins without involvement of the calf veins. The risk of embolisation posed by these thrombi is significant.

Clinical manifestations include swelling, feeling of heaviness, warmth and tenderness if the DVT is in the calf. These symptoms and signs, especially the swelling, can be present in normal pregnancy, and may be absent in isolated proximal DVT. Lower abdominal and groin pain, mild pyrexia, mild leukocytosis and swelling of the whole lower limb should raise the clinical suspicion of proximal thrombosis. It has been shown that experienced clinicians’ subjective prediction of DVT is good, and that three variables contribute significantly to this prediction (the LEFt rule): symptoms in the left leg (L), calf circumference difference greater or equal to 2 cm (E), and first trimester presentation (Ft).

The investigation of choice is compression ultrasound (CUS), preferably with Doppler (duplex ultrasound). The method is non-invasive, readily available, does not involve radiation and has good diagnostic performance. The sensitivity is 97%, and the specificity is 94%. The method is relatively insensitive for calf-vein thrombosis, but these rarely embolise. In cases of high clinical suspicion, where the CUS is negative, treatment should be instituted on clinical grounds, and either the CUS repeated in 1 week, or an alternative imaging method used.

The alternative method of choice is magnetic resonance venography (MRV). MRV can be carried out without intravenous contrast, and is highly superior in the assessment of thrombosis in the vena cava, pelvic veins and lower extremities. The sensitivity is 100% in the pelvis and thigh and 87% in the calf, with a specificity of 95–100%. To date, no harmful effect has been shown at any stage in pregnancy. Nevertheless, the current UK guidelines recommend that magnetic resonance imaging is not advisable in the first trimester, but it is preferable to ionising radiation. The major limitation of this imaging modality will be its local availability.

Pulmonary embolism

The most common clinical manifestations of pulmonary embolism are in decreasing frequency order (from 70% to 10%) dyspnoea, tachypnoea, chest pain, apprehension, tachycardia, cough with haemoptysis. Hypoxaemia and haemodynamic collapse are less common. Difficulties arise from the overlap of these symptoms with those of normal pregnancy, and the differentiation is often problematic. There are no hard and fast rules, but the occurrence of new chest symptoms, particularly if of sudden onset in a pregnant woman with risk factors, should always prompt investigations. Clinicians should use their clinical judgement and pursue diagnostic imaging for suspected pulmonary embolism.

The definitive imaging methods are associated with radiation exposure. The radiation to the mother and fetus ( Table 2 ) should not be a deterrent, but the basis of careful risk-benefit considerations. The choice of imaging test will also be dependent on the clinical suspicion and local availability, and should be made by discussion with the radiologist. Generally, the risk from missing such an important diagnosis far outweighs the risk to the mother and fetus from the investigations. A proposed algorithm is included below ( Fig. 1 ).

Proposed algorithm for the diagnosis and treatment of PE in pregnancy.

In individuals who are haemodynamically compromised with suspected pulmonary embolism, emergent bedside echocardiography is a useful adjunct. A massive pulmonary embolism will cause right ventricular enlargement and systolic dysfunction. Conversely, the absence of these makes pulmonary embolism as the cause of haemodynamic compromise unlikely.

The first investigation in a stable individual is a chest X ray. The role of the chest X ray is not diagnostic and will be normal in over 50% of cases. Its role is to detect alternative pathology, which may explain the symptoms and make an alternative diagnosis to help inform the definitive diagnostic imaging test. The fetal radiation dose, especially with abdominal shielding from a chest X ray, is negligible.

Many authorities have proposed CUS as the next step, and this should certainly be undertaken if there are signs of lower limb DVT. A diagnosis of DVT indicates full anticoagulation and obviates the need for further imaging and radiation. A proximal thrombus is found in 23–52% of pregnant women with confirmed pulmonary embolism, most of whom will have symptoms. In the absence of symptoms, at least outside pregnancy, CUS is not recommended, as it is often negative: 30% of women with negative CUS will have a pulmonary embolism.

If the chest X ray is normal, the lung perfusion element of a ventilation–perfusion (V–Q) scan should be preformed next. Seventy per cent of pregnant women had negative scans in a retrospective study, and the rate of recurrent events in this group was reassuringly low ; therefore, they were true negatives. If the perfusion scan is normal, the ventilation component can be omitted, thereby significantly reducing the risk from radiation. There are two main problems with V–Q scanning. The first concerns the high proportion of indeterminate results, up to 21% in older studies, which would imply that many women would have to undergo a further test with added radiation risk. More recent studies have shown that this is not the case if the chest X ray is normal, in which case the proportion of non-diagnostic V–Q scans is about 1.3–6%. The second concern relates to the increased risk of radiation to the fetus ( Table 2 ). It was estimated that the added risk of childhood cancer after V–Q (both perfusion and ventilation component) scan is 1 out of 280,000, compared with 1 out of 1,000,000 for computed tomography pulmonary angiography (CTPA). Interpretation of these estimates should also take into account the background risk of childhood cancer, which is 140 out of 1,000,000 per year in the UK (from Cancer research UK for childhood cancer: http://info.cancerresearchuk.org/cancerstats/childhoodcancer/incidence ).

If the chest X ray is abnormal, or there is unavailability or contraindication to the V–Q scan, and in those cases when the V–Q scan is non-diagnostic, CTPA should be carried out. This is the first-line investigation outside pregnancy owing to its higher sensitivity and specificity and smaller proportion of equivocal results. In addition, it can offer an alternative diagnosis, including serious and potentially fatal aortic dissection. In pregnancy, it seems that the diagnostic performance is not necessarily superior. The radiation to the fetus is lower than with the V–Q scan ; however, the radiation dose delivered to the mother’s breasts is significantly higher. This results in an estimated increase in the life-time risk of breast cancer, which is dependent on the age at the time of the computed tomography scan, one in 143 at age 20 years and one in 284 at the age of 40 years. These estimates are for non-pregnant women undergoing computed tomography scanning; they may be higher for the pregnant, lactating breast, or both. In view of these concerns, most authorities conclude that it should remain the second-line imaging in pregnancy and puerperium. Again, when weighing up these risks, it must be born in mind that the one in 8 women will develop breast cancer in their life-time (Cancer Research UK).

Cerebral vein thrombosis

Clinical features include, in reducing order of frequency, the following: a severe headache in 72%; seizures that can be partial, complex or generalised in 45%; confusion and altered consciousness in 43%; and signs of increased intracranial pressure (e.g. papilloedema, vomiting, photophobia) in 30% of women. One-third to two-thirds of women may have hemiparesis or hemineglect, but focal signs may be absent. Again, it must be remembered that CVT can occur from early pregnancy. The diagnosis is confirmed by imaging, and MRV is the investigation of choice. If this is not available, computed tomography is acceptable, but it must be contrast-enhanced (computed tomography venogram).