Chapter 41Thromboembolic Disease in Pregnancy Charles J. Lockwood, MDVenous thromboembolism (VTE) is the leading cause of maternal centrations at term.5 Thus, pregnancy is associated with increasedmortality in the United States, accounting for almost 20% of preg- thrombin-generating potential, decreased endogenous anticoagulantnancy-related deaths in the past decade.1 A retrospective cohort study effects, and impaired ﬁbrinolysis.of 268,525 patients over a 19-year period reported a prevalence of VTE The occurrence of VTE in pregnancy is also promoted by venousof 1 per 1627 births; of these cases, 77% were deep venous thromboses stasis in the lower extremities resulting from compression of the infe-(DVTs), and 23% were acute pulmonary emboli (PE).2 No antecedent rior vena cava and pelvic veins by the enlarging uterus, compressionhistory of VTE was present in 86% of these patients. Moreover, among of the left common iliac vein by the right iliac artery,8 and increases innonpregnant adults who have a fatal PE, 65% (95% conﬁdence inter- deep vein capacitance caused by increased circulating levels of proges-vals [CI], 40.8% to 84.6%) die within 1 hour after onset.3 These ﬁnd- terone and local endothelial production of prostacyclin and nitricings underscore the need for a high index of suspicion, a sensitive and oxide.9,10rapid diagnostic algorithm, and expeditious initiation of treatment inpregnant women with suspected VTE. Among pregnant women, 98.4% of DVTs are localized to the lower Risk Factors Not Speciﬁcextremities, with the left leg affected in 82% of cases.2 The occurrence to Pregnancyof DVT is more common in the antepartum than in the postpartum Additional risk factors for VTE that may be more common in preg-period (74% versus 26%; P < .001), with a mean gestational age at nancy include trauma, infection, obesity, severe proteinuria, and pro-diagnosis of 16.8 ± 2.4 weeks. Nearly 50% of antepartum DVTs are longed bed rest. Maternal age greater than 35 years doubles the risk ofdetected by 15 weeks, 38% between 16 and 30 weeks, and only 12% VTE in pregnancy.11 One study found that, among patients undergo-after 30 weeks. In contrast, most PEs are diagnosed in the postpartum ing cesarean delivery who developed a PE, 36% were older thanperiod (60.5%) and are strongly associated with cesarean delivery 35 years of age, and 55% were obese (body mass index >29).12(relative risk [RR], 30.3; P < .001).2 Antiphospholipid antibody (APA) syndrome is associated with a 1% to 5% risk of VTE in pregnancy and the puerperium despite thromboprophylaxis.13,14 In a case-control study of 30 pregnant women with VTE versus matched controls who were subsequently analyzedRisk Factors for APA, the prevalence of these antibodies was substantially increased in cases compared with controls (27% versus 3%; P = .026).15Pregnancy Is a Prothrombotic State The presence of an inherited thrombophilic disorder also increasesNormally, VTE is a disease of aging, occurring in fewer than 1 of every the risk of VTE during pregnancy, particularly in the setting of a per-10,000 healthy women before 40 years of age.4 However, the risk of sonal or strong family history. For example, the factor V Leiden (FVL)VTE is increased sixfold in pregnancy. Pregnancy induces this pro- mutation is present in 40% of pregnant patients with VTE.16,17 However,thrombotic state in a number of ways. Compared to nonpregnant because the prevalence of VTE in pregnancy is low (1/1600) and thewomen of reproductive age, pregnancy is associated with increases of incidence of heterozygosity for FVL in European populations is high20% to 1000% in plasma concentrations of ﬁbrinogen; factors VII, (5%), the actual risk of VTE among gravidas who are without a per-VIII, IX, X, and XII; and von Willebrand factor.5 In addition, activity sonal history of VTE or an affected ﬁrst-degree relative is less thanof the anticoagulant factor, protein S, declines, on average, to 39% of 0.2% to 0.3%.16,17 With such a history, the risk of VTE in the antepar-normal in the second trimester and 31% of normal in the third trimes- tum or postpartum period is greater than 10%. Similar observationster.6 As a consequence, pregnancy is associated with an increase in have been made for the other common inherited thrombophilias (seeresistance to activated protein C. The net effect of these changes is an Chapter 40 and Table 41-1).increase in thrombin generation, as measured by increased levels of The presence of a thrombophilia can also affect the recurrence riskﬁbrinopeptide A and the thrombin-antithrombin complex.7 Protein S for VTE among pregnant women. Brill-Edwards and colleagues pro-levels drop even further after cesarean delivery or infection, helping to spectively followed 125 pregnant women with a prior VTE, 95 ofaccount for the high prevalence of PE after cesarean deliveries. Levels whom were tested for thrombophilias, including FVL; the prothrom-of plasminogen activator inhibitor 1 (PAI-1), which inhibits clot lysis, bin G20210A gene mutation (PGM); protein C, protein S, and anti-increase threefold to fourfold during pregnancy, whereas plasma thrombin deﬁciencies; and the APAs, anticardiolipin antibodies, andPAI-2 values, which are negligible before pregnancy, reach high con- lupus anticoagulant.23 The authors withheld antepartum thrombopro-
856 CHAPTER 41 Thromboembolic Disease in Pregnancy TABLE 41-1 INHERITED THROMBOPHILIAS AND THEIR ASSOCIATION WITH VENOUS THROMBOEMBOLISM IN PREGNANCY Probability of VTE Probability of VTE % Thrombophilia Relative Risk in Patients without with a Personal or % VTE in in European or Odds Ratio Personal or Family Strong Family Thrombophilia Pregnancy Populations (95% CI) History of VTE (%) History of VTE (%) References FVL (homozygous) <1 0.06+ 25.4 [8.8-66] 1.5 17 16-19 FVL (heterozygous) 40-44 5 6.9 [3.3-15.2] 0.26 10 16,17 PGM (homozygous) <1 0.02+ NA 2.8 >17 18 PGM (heterozygous) 17 3 9.5 [2.1-66.7] 0.37 >10 16,17 FVL/PGM (compound <1 0.15 84 [19-369] 4.7 NA 16,17 heterozygous) Antithrombin 1-8 0.04 119 3.0-7.2 >40 16-21 deﬁciency Protein S deﬁciency 12.4 0.03-0.13 2.4 [0.8-7.9] <1 6.6 16-18, 21 Protein C deﬁciency <10 0.2-0.3 6.5-12.5 0.8-1.7 NA 16,18, 22 +, calculated based on a Hardy-Weinberg equilibrium; CI, conﬁdence interval; FVL, factor V Leiden; NA, not applicable; PGM, prothrombin G20210A gene mutation; VTE, venous thromboembolism.phylaxis but employed it in the postpartum period. They noted an TABLE 41-2 DVT CLINICAL CHARACTERISTICoverall antepartum recurrence rate of 2.4% (CI, 0.2% to 6.9%) but no SCORErecurrences in the 44 women without a detectable thrombophiliawhose previous VTE was associated with a temporary risk factor Item No. Description Score(among which the authors included pregnancy itself). In contrast, the 1 Immobilization due to cast or paresis +1recurrence risk for VTE among the 25 thrombophilic patients was 16% 2 Bed rest for >3 days or major surgery +1(4 patients) (odds ratio [OR], 6.5; CI, 0.8 to 56.3). Therefore, it would within 12 wk requiring general or regionalappear prudent to test pregnant patients who have a history of a VTE anesthesiaassociated with a transient risk factor (e.g., fracture) for thrombophil- 3 Localized tenderness along deep venous +1ias. Similarly, consideration should be given to screening pregnant systemwomen who have a strong family history (i.e., affected ﬁrst-degree 4 Entire leg swollen +1relative) of VTE, particularly if they are likely to be exposed to other 5 Asymmetric calf swelling >3 cm measured +1risk factors such as prolonged mobilization or cesarean delivery. 10 cm below tibial tuberosity 6 Pitting edema only in symptomatic leg +1 7 Collateral nonvaricose superﬁcial veins +1 +1Diagnosis and Evaluation of 8 9 Active cancer Prior documented DVT +1Venous Thromboembolism 10 Alternative diagnosis at least as likely as DVT −2in Pregnancy *Patients with a score <2 are considered unlikely, and those with a score ≥2 are considered likely, to have deep venous thrombosis (DVT).The clinical signs and symptoms of VTE are neither sensitive nor spe- Adapted from Wells PS, Anderson DR, Rodger M, et al: Evaluation ofciﬁc. Indeed, three quarters of patients in whom the diagnosis of either D-dimer in the diagnosis of suspected deep-vein thrombosis. N Engl JDVT or PE is suspected are unaffected.24 Conversely, many of those Med 349:1227-1235, 2003.ultimately diagnosed with VTE do not have classic features. To makean early diagnosis, clinicians must exercise a high index of suspicionand approach the diagnosis in a systematic fashion. Wells and associates introduced and validated a simple model in non- pregnant patients (Table 41-2).26 In the Wells model, 46% of nonpreg- nant patients were categorized as likely to have a DVT (score = 2), and,Deep Venous Thrombosis of these patients, 28% (CI, 24% to 32%) were found to have either proximal DVT or PE. In contrast, of the 54% who were categorized asClinical Presentation unlikely to have a DVT, only 6% (CI, 4% to 8%) proved to have aOnly a third of patients with unilateral lower extremity edema, ery- proximal DVT or PE. Therefore, the ﬁrst step in diagnosing DVT isthema, warmth, pain, tenderness, and a positive Homan sign—the risk ascertainment.traditional hallmarks of DVT—prove to have the diagnosis whenobjective diagnostic tests are performed.25 Differential diagnoses Venous Ultrasonographyinclude a ruptured or strained muscle or tendon, cellulitis, knee joint Venous ultrasonography (VUS) with or without color Doppler imaginginjury, Baker cyst, cutaneous vasculitis, superﬁcial thrombophlebitis, has become the primary diagnostic modality for evaluating pregnantand lymphedema. The positive predictive value of these signs and patients who are at risk for DVT. The test is performed by placing thesymptoms increases substantially in patients at increased risk. However, ultrasound transducer over the common femoral vein, beginning atthere is no risk assessment model that has been validated in pregnancy. the inguinal ligament, and then moving down the leg to sequentially
CHAPTER 41 Thromboembolic Disease in Pregnancy 857image the greater saphenous vein, the superﬁcial femoral vein, and mography (IPG), two clinically useful patterns emerged in 70% ofthen the popliteal vein to its trifurcation with the deep veins of the calf. at-risk patients.33 The ﬁrst pattern was a normal IPG and a negativeCalf veins are then insonated. Pressure is applied with the probe to D-dimer assay, which had a negative predictive value for DVT of 97%determine whether the vein under examination is compressible. The overall and 99% for proximal vein DVT. The second pattern was themost accurate ultrasonic criterion for diagnosing DVT is noncom- combination of a positive D-dimer and an abnormal IPG, which hadpressibility of the venous lumen in a transverse plane under gentle a positive predictive value of 93% for any DVT and 90% for proximalprobe pressure using duplex and color ﬂow Doppler.25 The sensitivity DVT. However, if the D-dimer and IPG results were discordant, it wasand speciﬁcity of VUS are reported to be 90% to 100% for proximal not possible to reliably exclude or diagnose DVT, and such discordantvein thromboses.27 Meta-analysis suggests that VUS is also effective at results occurred in 28% of patients. Extrapolation of these ﬁndings toscreening for calf vein DVT, with a sensitivity of 92.5% (CI, 81.8% to pregnancy is difﬁcult, not only because of the high “false-positive” rate97.9%) and a speciﬁcity of 98.7% (CI, 95.5% to 99.9%), yielding an of D-dimer testing (>50%)34 but because IPG is rarely performedoverall accuracy of 97.2% (CI, 93.9% to 99.0%).28 Even though there anymore.is a paucity of data on the performance of VUS in pregnancy, it has A number of investigators have argued that, given the high rate ofbecome the gold standard for DVT detection in the nonpregnant false-positive results on D-dimer testing in pregnancy, it can be usedstate.29 as a screening test in low-risk patients, because a negative D-dimer result in a pregnant woman would have an exceedingly high negativeMagnetic Resonance Venography predictive value. Morse measured D-dimer values in 48 women, agedIt appears that magnetic resonance (MR) imaging may be superior to 17 to 36 years, at 16, 26, and 34 weeks of gestation and compared theseVUS, and perhaps the equivalent of contrast venography, for diagnos- values to those of 34 healthy, nonpregnant controls.34 Morse found aing DVT. The sensitivity of MR imaging for the diagnosis of proximal progressive increase in D-dimer concentrations across gestation (191leg vein DVT has been reported to be 100% (CI, 87% to 100%); the ± 25 ng/mL at 16 weeks, 393 ± 72 ng/mL at 26 weeks, and 544 ± 96 ng/speciﬁcity, 100% (CI, 92% to 100%); and the overall accuracy, 96% mL at 34 weeks), all of which were signiﬁcantly higher than the values(CI, 89% to 99%).30 Moreover, MR venography is signiﬁcantly more in nonpregnant women (140 ± 58 ng/mL). Because the cutoff value forsensitive and more accurate than sonography for the detection of normal D-dimer levels at the author’s hospital was 280 ng/mL, mostpelvic and calf DVT. The published literature suggests that the range pregnant patients would be considered positive. The author recom-of sensitivity for MR imaging in the diagnosis of DVT is 80% to 100%, mended new threshold ranges for 16 to 26 weeks (<465 ng/mL) andand its speciﬁcity is 90% to 100%, with median published rates of for 27 to 34 weeks (<640 ng/mL) of gestation.100% for both.31 There is preliminary evidence that the combination of VUS and D-dimer values measured by the SimpliRED D-dimer test may beContrast Venography useful in diagnosing DVT in pregnancy.35 Based on the initial VUS andBefore improvements in sonographic imaging technology and the D-dimer results, women were categorized into one of four groups andintroduction of MR venography, contrast venography was the gold managed accordingly. Group 1 (VUS normal, D-dimer normal) com-standard for the diagnosis of DVT. The procedure involved injecting a prised 31 patients who had routine follow-up until 6 weeks postpar-contrast agent into a superﬁcial vein on the dorsum of the foot and tum; none of these patients developed objectively diagnosed VTE (CI,allowing it to circulate into the deep venous system while radiographic 0% to 9.2%). In group 2 (VUS normal, D-dimer abnormal), 4 (22%)images were obtained of the lower leg, thigh, and pelvis. The diagnosis of 18 women were diagnosed with DVT on serial VUS at 3 and 7 days,required intraluminal ﬁlling defects observed on two or more views or and, of the 14 (78%) with negative serial VUS, none developed VTEan abrupt cutoff of contrast material. It was the most sensitive test for during follow-up (CI, 0% to 19.3%). In patients for whom VUS wascalf vein DVTs. Although accurate, it was expensive, invasive, and equivocal (group 3), venography was performed. If the VUS result waspainful and risked radiation exposure. In addition, the contrast positive (group 4), DVT was diagnosed. The authors concluded thatmedium could potentially induce renal compromise and chemical the ﬁnding of a normal VUS together with a normal D-dimer resultphlebitis. or an abnormal D-dimer result coupled with reassuring serial VUS ﬁndings allows the safe exclusion of DVT in pregnant patients, but thatD-dimer Assays larger conﬁrmatory studies were needed.Laboratory evaluation of D-dimer concentrations has been advocatedas an exclusionary test for DVT in low-risk nonpregnant women. A Testing Algorithms for Deep VenousD-dimer study is very likely to be positive in a patient with DVT, but Thrombosis in Pregnancyit is also commonly positive in patients with uncomplicated pregnan- The cornerstone of the evaluation of pregnant patients for possiblecies. Conversely, if a D-dimer assay is negative in a pregnant patient DVT is a VUS. However, D-dimer assessment can be employed eitherwith suspected DVT, the diagnosis is even less likely than with a com- as an initial screen in low-risk patients or as an adjunct to VUS. Figureparable negative ﬁnding in the nonpregnant state. Recently developed 41-1 outlines a diagnostic paradigm which assumes the availability ofD-dimer assays include two rapid enzyme-linked immunosorbent a sensitive D-dimer assay and clinical risk assessment. In this para-assays (ELISAs) (Instant-IA D-Dimer, Stago, Asnières, France, and digm, low-risk patients are given a D-dimer test. If the result is nega-VIDAS DD, bioMérieux, Marcy-l’Etoile, France) and a rapid whole tive, they are discharged for routine follow-up. If the D-dimer test isblood assay (SimpliRED D-Dimer, Agen Biomedical, Brisbane, Austra- positive, they undergo a VUS. Patients with moderate or high risk onlia). In nonpregnant patients, the sensitivity of the rapid ELISAs is clinical assessment proceed directly to VUS. If the VUS result is posi-greater than 95% and that of the SimpliRED D-dimer assay is report- tive, DVT is diagnosed and the patient is treated. If it is negative andedly 85%.32 In nonpregnant patients, there appears to be utility in the patient remains symptomatic, serial VUS testing can be performedcombining a sensitive D-dimer assay with a noninvasive imaging test. in 3 days, 7 days, or both. If the repeat VUS ﬁndings are positive, theWells and associates reported that, when the whole blood assay for D- patient is treated. If both VUS studies are negative, the patient hasdimer (SimpliRED) was used in combination with impedance plethys- routine follow-up. In particularly high-risk settings (e.g., known
858 CHAPTER 41 Thromboembolic Disease in Pregnancy Assess patient risk VUS Low Moderate Positive Negative or high Treat Assess Assess ( ) VUS for DVT risk D-dimer ( ) ( ) ( ) High Low ( ) D/C Consider serial VUS, Treat patient MR venogram, or ( ) ( ) contrast venography ( ) Serial VUS, Discharge MR venogram, orFIGURE 41-1 Testing algorithm for deep venous thrombosis contrast venographyin pregnancy, assuming availability of D-dimer test and clinicalassessment. D/C, discharge patient; MR, magnetic resonance;VUS, venous ultrasonography. FIGURE 41-2 Testing algorithm for deep venous thrombosis (DVT) in pregnancy with venous ultrasonography (VUS) as the initial test. MR, magnetic resonance.thrombophilia, suspected iliac vein thrombus), more deﬁnitive diag-nostic tests can be performed, such as an MR or contrast venogram ifthe D-dimer is positive and the initial VUS is negative. The problem Assess VUS and D-dimerwith this approach is that, as noted previously, the majority of preg-nant women after 16 weeks’ gestation will have a positive D-dimerassay if a nonpregnant threshold is used. Moreover, there has been no VUS ( ) and VUS ( )validation of higher D-dimer thresholds in pregnancy, and there are D-dimer ( )no validated risk scoring systems. Importantly, the Wells criteria may VUS ( ) butnot be appropriate in pregnancy, which is intrinsically a high-risk D-dimer ( )state. ( ) D/C patient Figure 41-2 assumes the use of a VUS as the initial test, with patientssubsequently triaged based on the VUS result or clinical risk assess-ment determining the need for additional evaluations. A positive VUS ( ) Consider serial VUS, ( ) MR venogram, or Treatresult prompts treatment. Follow-up of a negative VUS result depends contrast venographyon the patient’s risk category. If it is low, the patient is discharged toroutine follow-up. If it is moderate or high, she undergoes serial VUS FIGURE 41-3 Testing algorithm for deep venous thrombosis inor, if the index of suspicion is high enough, MR or contrast pregnancy, using venous ultrasonography (VUS) and D-dimervenography. testing without formal clinical risk assessment. D/C, discharge; The diagnostic scheme in Figure 41-3 utilizes VUS and D-dimer MR, magnetic resonance.testing without formal clinical risk assessment to triage patients. Thecombination of a negative D-dimer result and a negative VUS study isassociated with a very low risk of DVT, and such patients are dis- negative D-dimer ﬁnding would be very reassuring, whereas observa-charged to routine follow-up. A positive VUS, which is virtually always tion of an intraluminal ﬁlling defect on MR or contrast venographyassociated with a positive D-dimer test, prompts treatment. However, would be diagnostic of recurrent DVT.32a negative VUS result combined with a positive D-dimer test is fol-lowed with either serial VUS or MR or contrast venography, dependingon the clinician’s index of suspicion. It should be noted that D-dimer Acute Pulmonary Embolustesting is likely to be completely irrelevant in puerperal and postopera-tive patients, because these patients have very high false-positive Clinical Findingsrates.36,37 The diagnostic approach outlined in Figure 41-2 should be In nonpregnant women, the classic presentation of PE includes tachy-employed in those settings. pnea (>20 breaths/min) and tachycardia (>100 beats/min), either of The diagnosis of recurrent DVT in pregnancy presents a diagnostic which is present in 90% of cases.39 Additional common symptomschallenge, because VUS ﬁndings remain abnormal after an initial include dyspnea and pleuritic chest pain. In contrast, lightheadednessthrombus for up to 1 year in as many as 50% of patients.32 In this and syncope are rare and are indicative of massive emboli.40 In thesetting, an increase of more than 4 mm in the compressed diameter of nonpregnant state, the nonspeciﬁc nature of these signs and symptomsa previously involved vein has been reported to provide strong evi- is a result of the common comorbidities found in older patients, includ-dence of recurrent thrombosis, but this observation requires conﬁrma- ing viral and bacterial pneumonia, postoperative atelectasis, pneumo-tion.38 In such cases, strong consideration should be given to D-dimer thorax, exacerbation of chronic obstructive lung disease, congestiveassessment and adjunct imaging with MR or contrast venography. A heart failure, lung cancer, musculoskeletal chest wall pain, esophageal
CHAPTER 41 Thromboembolic Disease in Pregnancy 859 TABLE 41-3 PULMONARY EMBOLISM RISK right bundle branch block, P-wave pulmonale, or right axis deviation) SCORE* and therefore are rare.42 Moreover, such changes are not universal, even among critically ill patients. Constantini and associates reported Scoring Factor Points changes characteristic of right ventricular strain including an S1Q3 pattern in 67% of such patients, with a “septal embolic pattern” in 53%, Clinical signs and symptoms of DVT +3.0 and either anterior lead T-wave inversion or new right bundle branch Alternative diagnosis deemed less likely than PE +3.0 Heart rate >100 beats/min +1.5 block in only 16%.43 However, there are no data on ECG changes in Immobilization or surgery in previous 4 wk +1.5 pregnant women with PE. Moreover, pregnancy-induced physiologic Prior VTE +1.5 changes mimic left heart strain (e.g., T inversions in the left precordial Hemoptysis +1.0 leads) and, therefore, may mask PE-induced right heart strain changes. Active cancer +1.0 Conversely, some physiologic changes may simulate PE changes (e.g., physiologic Q waves in leads 3 and aVF), leading to false-positive *Clinical probability of PE is low with a cumulative score <2.0, results. For these reasons, the screening value of ECG in the setting intermediate with a score of 2.0-6.0, and high with a score >6.0. DVT, deep venous thrombosis; PE, pulmonary embolism; VTE, venous of maternal acute PE is probably low. However, assessment of the thromboembolism. maternal ECG may have utility in making decisions about how Adapted from Fedullo PF, Tapson VF: Clinical practice: The evaluation aggressively to pursue secondary diagnostic studies (see Testing of suspected pulmonary embolism. N Engl J Med 349:1247-1256, 2003. Algorithms for Pulmonary Embolism, later). ARTERIAL BLOOD GASESspasm, pericarditis, pleuritis, and anxiety. Although these comorbidi- In young, nonpregnant patients, assessments of arterial blood gasesties are far less common in otherwise young and healthy pregnant and oxygen saturation are of limited value in the setting of acute PE,women, pregnancy itself can be associated with dyspnea, tachycardia, because PO2 values higher than 80 mm Hg are found in 29% of suchlightheadedness, orthostatic presyncope, and various chest wall com- patients who are younger than 40 years of age.44 Given that PO2 levelsplaints. Therefore, a high index of suspicion is required. are even higher in pregnant women, the test is likely to produce even The initial step in the evaluation of any patient is risk assessment. more false-negative results in pregnancy. However, as was the case withTable 41-3 outlines one scoring system used in nonpregnant patients, maternal ECG, assessment of maternal oxygen saturation may havein which patients in the low-probability group had a prevalence of PE utility in making decisions about how aggressively to pursue secondaryof 10% or less, those in the intermediate group had a prevalence of diagnostic studies (see Testing Algorithms for Pulmonary Embolism,about 30%, and those in the high-probability group had a prevalence later).of PE of 70% or more.40 This system has not been validated in pregnantpatients. Furthermore, given the thrombogenic nature of pregnancy ECHOCARDIOGRAPHYand the puerperal state, particularly after cesarean delivery, it may be More than 80% of nonpregnant patients with acute PE have echo-prudent to consider all pregnant patients as being at high risk. cardiographic or Doppler abnormalities, including dilated and hypo- kinetic right ventricle and tricuspid regurgitation.45,46 Because young,Nonspeciﬁc Diagnostic Tests pregnant women are unlikely to have cardiorespiratory comorbidities CHEST RADIOGRAPHY that can mimic these ﬁndings (e.g., chronic obstructive pulmonary Chest radiographs can play an important role in the evaluation of disease), maternal echocardiography may be useful in pregnant womenpregnant patients with suspected PE. Given concerns about excess with suspected PE. Optimal screening requires transesophageal echo-breast irradiation during computed tomographic pulmonary angiog- cardiography with or without contrast enhancement.47 Indeed, there israphy (CTPA), and because ventilation-perfusion (V/Q) studies have a report in which intraoperative transesophageal echocardiographyexceptionally high negative predictive values in otherwise healthy was used to diagnose a right atrial thrombus during a cesareanpregnant women, some investigators have advocated obtaining an delivery.48initial chest radiograph to triage patients to either CTPA or V/Q scan.41If the radiograph is normal, a V/Q scan is obtained; if it is abnormal Speciﬁc Diagnostic Tests(e.g., inﬁltrates, which could compromise the V/Q study by causing VENTILATION-PERFUSION SCANNINGmatched ventilation-perfusion abnormalities), a CTPA is ordered. In nonpregnant adults, V/Q scanning has largely been replaced byAbnormal ﬁndings on radiography may also be consistent with a PE. CTPA. In younger, otherwise healthy, pregnant women, V/Q scans haveSuch ﬁndings, which include pleural effusion, pulmonary inﬁltrates, superior PE diagnostic characteristics compared with their use in older,atelectasis, and elevated hemidiaphragm, underscore the need for nongravid patients, who frequently have cardiorespiratory comorbidi-CTPA, because it can diagnose a broad range of lung pathology. ties.41 Because of this improved accuracy in pregnancy, and given con-However, chest radiography is rarely directly diagnostic of a PE, because cerns about the extent of maternal breast irradiation with CTPA, V/Qthe classic radiographic ﬁndings of pulmonary infarction, such as scans may still have a limited role in ruling out PE in select pregnantHampton’s hump or decreased vascularity (Westermark sign), are and puerperal patients.41 The perfusion component of a V/Q scanrarely seen.31 requires injecting human albumin macroaggregates labeled with radioactive isotopes (e.g., technetium 99m) into the bloodstream, ELECTROCARDIOGRAPHY where they are deposited in the pulmonary capillary bed and imaged In nonpregnant patients, electrocardiographic (ECG) changes are by a photoscanner. The ventilation component requires inhalation ofpresent in 87% of patients with documented PE who are otherwise radiolabeled (e.g., xenon 133) aerosols whose distribution in the alveo-without underlying cardiopulmonary disease.42 However, truly charac- lar space is assessed by a gamma camera. Comparison of the perfusionteristic ECG changes usually reﬂect the hemodynamic sequelae of and ventilation scans produces characteristic patterns that can be usedacute cor pulmonale characteristic of a massive PE (i.e., S1Q3T3 pattern, to assign diagnostic probabilities.
860 CHAPTER 41 Thromboembolic Disease in Pregnancy The chief limitation of V/Q scanning in nonpregnant patients is contrast pulmonary angiogram and a negative (normal or near-that only 30% to 40% of older patients with preexisting lung disease normal) V/Q scan.who have large matched V/Q defects are subsequently shown to have It had been argued that CTPA is less accurate with small, isolateda PE, leading to false-positive results.25 Fortunately, pregnant patients subsegmental, peripheral vessels and horizontally oriented vessels inare generally free of chronic lung disease. Moreover, as noted earlier, the right middle lobe. However, newer technology using spiral com-the proportion of nondiagnostic V/Q scans can be further reduced by puted tomographic scanners with multiple detector rows and 1-mmobtaining an initial chest radiograph and triaging pregnant patients slices appears to improve the accuracy of CTPA for diagnosing subseg-with an abnormal result to CTPA.41 Furthermore, although fetal radia- mental infarcts and may reduce false-negative results to 5%.54 In addi-tion exposure is higher with V/Q scanning than with CTPA (0.11 to tion, traditional contrast pulmonary angiograms have relatively poor0.22 mGy versus 0.003 to 0.08 mGy),49 all of these exposures are interobserver agreement (45% to 66%) for diagnosis of subsegmentalexceedingly low, and the risk of subsequent leukemogenesis is esti- PE.55,56 Therefore, CTPA appears to be of comparable or superior efﬁ-mated at 1 case per 16,000 fetuses exposed to 1 mGy of radiation. cacy to contrast angiography.Moreover, half-dose perfusion scans can be employed in young, healthy,pregnant women. Such half-doses result in breast irradiation levels of MAGNETIC RESONANCE ARTERIOGRAPHY0.25 mGy, which is 140-fold lower than the exposure from CTPA (i.e., Meaney and colleagues compared standard conventional intrave-35 mGy per breast).41 nous contrast pulmonary angiography with magnetic resonance angi- The diagnostic efﬁcacy of V/Q scanning for PE was assessed by the ography (MRA) during the pulmonary arterial phase of the cardiacmulticenter Prospective Investigation of Pulmonary Embolism Diag- cycle after an intravenous bolus of gadolinium and observed that MRAnosis (PIOPED) in 931 nonpregnant adults.50 In this study, patients had an overall sensitivity of 100%, a speciﬁcity of 95% (CI, 87% towith high-probability scans were found by deﬁnitive testing (e.g., pul- 100%), and positive and negative predictive values of 87% (CI, 74%monary contrast angiography) to have PE in 87% of cases, but only to 100%) and 100%, respectively.57 Oudkerk and associates conducted41% of patients with PE had high-probability scans (sensitivity, 41%; a prospective study of MRA in which conventional contrast pulmonaryspeciﬁcity, 97%). In contrast, 33% of patients with intermediate- angiography results were available for 118 patients at risk for PE.58 Theprobability scans had a PE, and PE was present in only 14% of patients prevalence of PE was 30%, and MRA identiﬁed 27 of 35 patients withwith low-probability scans and 4% of those with negative scans. Inves- angiographically conﬁrmed PE (sensitivity, 77%; CI, 61% to 90%). Thetigators also found that PE can be present in a substantial percentage sensitivity of MRA for isolated subsegmental, segmental, and centralof high-risk patients with nondiagnostic V/Q scans. One study found or lobar PE was 40%, 84%, and 100%, respectively. In fact, MRA identi-that, among high-risk women, the prevalence of PE in those with low ﬁed PE in two patients with normal angiograms. Given its lack ofand intermediate probability on V/Q scanning was 40% and 66%, radiation exposure, and if its efﬁcacy is conﬁrmed by larger studies,respectively.31 Conversely, 44% of low-risk patients with a high- MRA may become the primary diagnostic test, where it is available, forprobability V/Q scan did not have a PE. These ﬁndings underscore the ruling out PE in pregnancy.value of ﬁrst determining the patient’s clinical risk category if V/Qscanning is to be used as a ﬁrst-line screening method. This is prob- CONVENTIONAL CONTRASTlematic in pregnancy, because there are no validated risk scoring PULMONARY ARTERIOGRAPHYsystems. Pulmonary arteriography was long considered the gold standard The efﬁcacy of V/Q scanning in pregnancy was assessed by Chan for the diagnosis of PE, with sensitivities and speciﬁcities of 100%and coworkers.51 They evaluated 120 consecutive pregnant women by deﬁnition. However, as noted earlier, interobserver agreementhaving 121 V/Q scans for suspected PE. All V/Q results were reinter- decreases with smaller peripheral lesions. To perform the study, apreted by two independent experts. Eight of the patients (6.6%) were catheter is usually placed in the right femoral, basilic, or right internalalready receiving treatment for VTE before their V/Q scan. Of the jugular vein. A PE is diagnosed by the ﬁnding of an intraluminal ﬁllingremaining 113 scans, 83 (73.5%) were interpreted as normal, 28 defect on two views of a pulmonary artery. The procedure has a 0.5%(24.8%) as nondiagnostic, and 2 (1.8%) as high-probability. None of mortality rate and a complication rate of 3%, including sequelae ofthe 104 women who did not receive treatment because of a reassuring contrast injections and catheter placement such as respiratory failurescan (i.e., 80 with normal and 24 with nondiagnostic results) had a (0.4%), renal failure (0.3%), cardiac perforation (1%), and groinsubsequent VTE over a follow-up period of more than 20 months. This hematoma requiring transfusion (0.2%). In light of these risks, pul-study suggested that the negative predictive value of V/Q scanning is monary arteriography has fallen out of favor given the relative safetyhigher in pregnancy than among older, nonpregnant patients. and efﬁcacy of the other available tests.40,59,60 Moreover, contrast pul- monary arteriography is relatively contraindicated in the presence of COMPUTED TOMOGRAPHIC a signiﬁcant hemorrhagic risk (e.g., disseminated intravascular coagu-PULMONARY ANGIOGRAPHY lation, thrombocytopenia) and in patients with renal insufﬁciency. Left CTPA has become the gold standard for the diagnosis of PE in bundle branch blocks require the use of a temporary pacemaker duringnonpregnant patients.41 Cross and colleagues compared CTPA to V/Q the procedure to avoid induction of complete heart block.31scans for the initial investigation of patients with suspected PE and Pulmonary arteriography from the brachial vein route generatesobserved that deﬁnitive diagnoses of various etiologies were more fre- 0.5 mGy (0.05 rad) of fetal exposure, whereas the femoral vein approachquent with CTPA (90% versus 54%; P < .001).52 However, although generates 2.2 to 3.3 mGy (0.22 to 0.33 rad) of fetal exposure, makingCTPA more often demonstrated nonembolic lesions responsible for the former the preferable route in pregnancy.61 Concerns also existthe patients’ symptoms, there was no difference in the detection rate regarding the development of fetal goiter after maternal exposure toof PE between the two groups. Meta-analysis of 23 studies demon- iodinated contrast material. Therefore, the fetal heart rate should bestrated very low 3-month rates of subsequent VTE (1.4%; CI, 1.1% to assessed biweekly to rule out hypothyroidism, and, if delivery occurs1.8%) and fatal PE (0.51%; CI, 0.33% to 0.76%) after a negative CTPA proximate to the test, neonatal thyroid function should be checkedstudy.53 These results compare very favorably with both a negative during the ﬁrst week after the procedure.62 Given the high sensitivity
CHAPTER 41 Thromboembolic Disease in Pregnancy 861and speciﬁcity of CTPA and MRA, as well as the relatively high mater-nal morbidity and concerns with both fetal irradiation and iodinated Assess risk scorecontrast exposure attendant on contrast pulmonary arteriography, thelatter has fallen into disuse as a tool for diagnosing PE in pregnancy. Low 2 Not low 2 EVALUATION OF LOWER EXTREMITIES FORVENOUS THROMBOSIS ( ) Approximately 75% of patients with PE have imaging evidence of VUS and CXRDVT, and two thirds of those lesions are present in the proximal vein. D-dimerHowever, only a quarter of patients with symptomatic PE have symp- ( ) ( )tomatic DVT.63 Therefore, in stable patients before deﬁnitive diagnos- ( )tic imaging, and in high-risk pregnant patients in whom CTPA, V/Q ( )scanning, or MRA is not diagnostic, detection of leg vein DVT by VUS ( ) D/C Tx CTPA V/Qcan establish the need for anticoagulation.64 ( ) ( ) ( ) D-DIMER ASSAYS AS A SCREEN FOR ( ) ( )PULMONARY EMBOLISM MRA, contrast Low or Tx D/C A negative D-dimer concentration (<500 ng/mL) has been advo- angiography, or interm.cated as a “rule-out” test in patients with low probability of PE because serial VUS prob.of its high (95%) negative predictive value.31 Kearon and colleaguesobserved that, among patients at low clinical risk for PE, none of those FIGURE 41-4 Testing algorithm for pulmonary emboli with initialwith a negative D-dimer result and no additional diagnostic testing triage based on clinical probability. CTPA, computed tomographichad a subsequent VTE, compared with 1 patient with VTE among 182 pulmonary angiography; CXR, chest radiography; D/C, discharge;with negative D-dimer result who had additional testing.65 They con- interm. prob., intermediate probability; MRA, magnetic resonancecluded that, for patients with a low probability of PE whose D-dimer angiography; Tx, treatment; V/Q, ventilation-perfusion studies; VUS,result is negative, additional diagnostic testing can be withheld without venous ultrasonography.increasing the frequency of VTE. Meta-analysis of studies examining the accuracy of D-dimer deter-minations in the diagnosis of PE demonstrate a mean sensitivity of95% (CI, 88.0% to 100%), a speciﬁcity of 45% (CI, 38% to 53%), and Assess D-dimer and CTPApositive and negative likelihood ratios of 1.74 (CI, 1.55 to 1.91) and0.11 (CI, 0.03 to 0.39), respectively.66 The quantitative rapid ELISA hada sensitivity of 98% (CI, 88.0% to 100%), a speciﬁcity of 40% (CI, 29% D-dimer ( ) D-dimer ( ) CTPA ( )to 50%), and positive and negative likelihood ratios of 1.62 (CI, 1.38 CTPA ( ) CTPA ( )to 1.91) and 0.05 (CI, 0.00 to 4.15), respectively. The whole blood D-dimer assay kit yielded a sensitivity of 82% (CI, 74 to 91), a speciﬁcityof 63% (CI, 54% to 71%), and positive and negative likelihood ratios ( ) ( )of 2.21 (CI, 1.81 to 2.70) and 0.28 (CI, 0.18 to 0.43), respectively. The ( ) D/C Consider serial VUS, Treatauthors concluded that a negative result on quantitative rapid ELISA patient MRA, or contrastwas as diagnostically useful as a normal V/Q scan for excluding PE. As angiographywas discussed for the use of D-dimer assays in the diagnosis of DVT,the speciﬁcity and positive predictive value for D-dimer assays in the FIGURE 41-5 Testing algorithm for pulmonary emboli withworkup of patients at risk for PE are likely to be far lower in pregnant, triage based on results from D-dimer testing and computedpuerperal, and postoperative patients, whereas the sensitivity and tomographic pulmonary angiography (CTPA). D/C, discharge;negative predictive value should be higher in these settings. MRA, magnetic resonance angiography; VUS, venous ultrasonography.Testing Algorithms for Pulmonary EmbolismThere is considerable controversy regarding the optimal paradigm fordiagnosing PE in pregnancy. Clinical risk assessment tools have not are deemed to be at low risk then undergo both VUS and D-dimerbeen validated for pregnant patients. D-dimer testing is likely to be testing; those at high risk and those with a positive D-dimer test receiveassociated with a very high false-positive rate. Moreover, as noted chest radiography followed by either CTPA or V/Q scanning. Thisearlier, concern has been expressed regarding the relatively high radia- approach would be favored in nonacute settings with equivocal signstion exposure to maternal breasts after CTPA. Conversely, V/Q scans and symptoms. The second approach (Fig. 41-5) seeks to maximize thedeliver higher radiation doses to the fetus, have limited application in speed and sensitivity of the diagnosis and should be used for patientsthe setting of cardiorespiratory comorbidities, and are frequently no who are more symptomatic or where there is a high index of suspicion.longer available in contemporary radiology practices. MRA may prove In this paradigm, triage is done with D-dimer testing and CTPA.to be the ideal alternative, but there are currently too few studies to rec- In Figure 41-4, patients should be minimally symptomatic andommend its widespread use, even where the technology is available. preferably should have reassuring oxygen saturation values (>80%). Two general strategies emerge. The ﬁrst (Fig. 41-4) seeks to mini- The workup commences with a risk assessment (see Table 41-3).mize radiation exposure to both the fetus and maternal breasts by Patients who are considered to be at low clinical risk (i.e., risk scoreperforming an initial triage based on clinical probability. Patients who <2) should have a VUS and D-dimer test. Obviously, if the VUS result
862 CHAPTER 41 Thromboembolic Disease in Pregnancyis positive, treatment should be initiated. If both the D-dimer and the TABLE 41-4 ADMINISTRATION OFVUS studies are negative, the patient can be discharged to routine INTRAVENOUS HEPARIN USING Afollow-up. If the VUS is negative but the D-dimer test is positive, the WEIGHT-BASED NOMOGRAM*patient is regarded as if she had a clinical risk assessment that was notlow (i.e., ≥2). In these latter patients, the workup commences with a aPTT Value Adjustmentchest radiograph. If the results are positive, the patient proceeds to a <35 sec (<1.2 × control) Repeat 80 U/kg bolus, then increaseCTPA. If this is positive for PE, treatment is immediately commenced. infusion rate by 4 U/kg/hrIf the CTPA is negative, consideration can be given to a serial VUS if 35-45 sec (1.2-1.5 × control) Repeat 40 U/kg bolus, then increasenot already done. If other ancillary tests are not reassuring (e.g., low infusion rate by 2 U/kg/hroxygen saturation, abnormal ECG), then consideration should be given 46-70 sec (1.6-2.3 × control) No change in dosingto an MRA or contrast angiogram. Alternatively, if the chest radio- 71-90 sec (2.4-3.0 × control) Decrease infusion rate by 2 U/kg/hrgraph is negative, a V/Q scan is performed. If this is negative or normal, >90 sec (>3.0 × control) Stop infusion for 1 hr, then restartthe patient can be discharged to routine follow-up. If the V/Q scan original dose decreased byindicates a high probability of PE, treatment should begin. If the V/Q 3 U/kg/hrscan returns an intermediate or low probability of PE, then considera- *Give bolus of 80 U/kg of body weight, followed by a maintenancetion can be given to a serial VUS if not already done, or, if other ancil- dose of 18 U/kg/hr. Assess aPTT values every 4-6 hr and makelary tests are not reassuring, an MRA or contrast angiogram can be done. adjustments made based on the aPTT values obtained. The paradigm in Figure 41-5 should be used for women with more aPTT, activated partial thromboplastin time.pronounced symptoms or unfavorable signs (e.g., oxygen saturation Adapted from Raschke RA, Reilly BM, Guidry JR, et al: The weight- based heparin dosing nomogram compared with a “standard care”values <80%, abnormal ECG). The workup begins with D-dimer nomogram: A randomized controlled trial. Ann Intern Med 119:874-determination and CTPA. In patients with a high pretest probability 881, 1993.of PE or who are very symptomatic, anticoagulation should be com-menced as soon as the blood sample is sent to the laboratory for D-dimer determination. If both the D-dimer and CTPA studies arenegative, the patient is at very low risk and can be discharged to routine The overall goal is to obtain and maintain an aPTT of 1.5 to 2.5 ×follow-up, though the cause of her symptoms must be identiﬁed (e.g., control values. The aPTT should not be used to guide unfractionatedpneumonia, pulmonary edema, cardiomyopathy, esophagitis). If the heparin therapy in patients with prolonged aPTT values due to theCTPA is positive, then anticoagulant treatment is begun or continued. presence of lupus anticoagulants. In these patients, plasma heparinIf the CTPA is negative and the D-dimer value is positive or the index activity can be measured by either a protamine sulfate or an anti-factorof suspicion remains high because of low oxygen saturation or an Xa chromogenic assay. Target plasma heparin concentrations of 0.2 toabnormal ECG, consideration should be given to a serial VUS, MRA, 0.4 U/mL are equivalent to anti-factor Xa concentrations of 0.4 toor contrast angiography. 0.7 U/mL. The usual duration of intravenous heparin therapy is 5 days, although patients with large iliofemoral thromboses or massive PEs should receive heparin for 7 to 10 days or until clinical improvement is noted.31 After hospital discharge, therapeutic doses of unfractionatedTreatment of Venous heparin are administered subcutaneously every 8 to 12 hours to main- tain the aPTT at 1.5 to 2 × control values 6 hours after the injection.Thromboembolism This therapy should be continued for 20 weeks and then followed byin Pregnancy prophylactic dosing until delivery. The standard prophylactic regimen of unfractionated heparin usedPatients with new-onset VTE during pregnancy should receive thera- in pregnancy consists of 5000 units administered subcutaneously everypeutic anticoagulation for at least 4 months during the pregnancy, 12 hours, increased by 2500 U in the second and third trimesters.followed by prophylactic therapy continuing for at least 6 weeks after However, Barbour and associates observed that this standard heparindelivery. During pregnancy, either unfractionated heparin or low- regimen was inadequate to achieve the desired anti-factor Xa thera-molecular-weight heparin (LMWH) is the anticoagulant of choice, peutic range in 5 of 9 second-trimester pregnancies and in 6 of 13given its efﬁcacy and safety proﬁle. Neither formulation crosses the third-trimester pregnancies.68 Therefore, careful assessment of anti-placenta, and neither poses teratogenic risks. After delivery, oral anti- factor Xa levels 4 to 6 hours after injection is required to properlycoagulation with warfarin may be started and is considered safe in adjust the dosage.breastfeeding mothers. The primary risks of long-term heparin therapy If vaginal or cesarean delivery occurs more than 4 hours after ain pregnancy are hemorrhage and osteoporosis. prophylactic dose of unfractionated heparin, the patient is not at sig- niﬁcant risk for hemorrhagic complications. Patients receiving unfrac- tionated heparin who experience bleeding or require rapid reversal ofHeparin Therapy the anticoagulant to effect delivery can be administered protamine sulfate by slow intravenous infusion of less than 20 mg/min, with noTherapeutic Unfractionated Heparin more than 50 mg given over 10 minutes. The amount of protamineThe initial intravenous unfractionated heparin dosage for pregnant needed to neutralize heparin is derived by determining the amount ofpatients with acute VTE should be determined with the use of a weight- residual heparin in the circulation, assuming a half-life for intrave-based nomogram, and subsequent dosage modiﬁcations should be nously administered heparin of 45 minutes. Full neutralization ofpredicated on the activated partial thromboplastin time (aPTT) (Table heparin activity would require 1 mg protamine sulfate per 100 units41-4). This regimen has been shown in nonpregnant patients to reduce of residual circulating heparin. If the heparin was administered sub-recurrence rates.67 cutaneously, repeated small infusions of protamine are required.
CHAPTER 41 Thromboembolic Disease in Pregnancy 863Finally, antithrombin concentrates may be used in antithrombin- without a personal or strong family history of VTE should receivedeﬁcient patients in the peripartum period. subtherapeutic doses of LMWH, with the goal of maintaining anti- One of the most serious potential complications of heparin therapy factor Xa levels of 0.3 to 0.7 U/mL 4 hours after injection.is heparin-induced thrombocytopenia (HIT). This condition arises in Regional anesthesia is contraindicated within 24 hours after thera-3% of nonpregnant patients given initial heparin therapy. Type 1 HIT peutic LMWH administration because of the risk of epidural hema-(HIT-1) occurs within days after initial heparin exposure, results from toma; therefore, we recommend switching to unfractionated heparinbenign platelet clumping in vitro, is self-limited, is not associated with at 36 weeks, or earlier if preterm delivery is expected. However, vaginala signiﬁcant risk of hemorrhage or thrombosis, and does not require or cesarean delivery occurring more than 12 hours after a prophylacticcessation of therapy. In contrast, type 2 HIT is a rare, immunoglobu- dose or 24 hours after a therapeutic dose of LMWH should not belin-mediated syndrome paradoxically associated with venous and arte- associated with treatment-induced hemorrhage. If shorter intervals arerial thrombosis that occurs 5 to 14 days after initiation of therapy. encountered, protamine may partially reverse the anticoagulant effectsMonitoring for HIT should include every-other-day platelet counts for of LMWH. The dosage is 1 mg of protamine for every 100 anti-Xa2 weeks.69 Because it can be difﬁcult to distinguish the two entities, a units of LMWH, but anti-factor Xa activity can be only partially (80%)50% decline in platelet count from its pretreatment high should reversed.73prompt cessation of therapy. The diagnosis of HIT-2 is conﬁrmed by The risk of HIT-2 appears to be far lower in patients receivingserotonin release assays, heparin-induced platelet aggregation assays, LMWH compared to unfractionated heparin, and it is lower still forﬂow cytometry, or solid phase immunoassays.70 If it is conﬁrmed, all obstetric patients receiving prophylactic LMWH therapy.74,75 However,forms of heparin, including intravenous ﬂushes, must be avoided. platelet counts should still be obtained every 2 or 3 days from day 4 to day 14.69Low-Molecular-Weight HeparinIn nonpregnant patients, the mainstay of acute treatment of VTE is Postpartum Anticoagulationnow LMWH. Meta-analyses of 22 studies including 8867 nonpregnant Unfractionated heparin or LMWH can be restarted 4 to 6 hours afterpatients suggest that LMWH has fewer thrombotic complications than vaginal delivery or 6 to 12 hours after cesarean delivery. Postpartumunfractionated heparin (3.6% versus 5.4%; OR, 0.68; CI, 0.55 to 0.84; patients should be started immediately on warfarin. The initial doses18 trials), produces a greater reduction in thrombus size (53% versus of warfarin should be 5 mg daily for 2 days. Subsequent doses are45%; OR, 0.69; CI, 0.59 to 0.81; 12 trials), and results in fewer major determined by monitoring the international normalized ratio (INR).hemorrhages (1.2% versus 2.0%; OR, 0.57; CI, 0.39 to 0.83; 19 trials) To avoid paradoxical thrombosis and skin necrosis from warfarin’sand fewer deaths (4.5% versus 6.0%; OR, 0.76; CI, 0.62 to 0.92; 18 initial anti-protein C effect, it is critical to maintain these women ontrials).71 However, comparable data have not been assembled for preg- therapeutic doses of unfractionated heparin for a minimum of 5 daysnant patients. Nevertheless, LMWH is now commonly employed to and until the INR has been at therapeutic levels (2.0 to 3.0) for 2 con-treat acute VTE in pregnancy and also as prophylaxis. secutive days. After an uncomplicated initial VTE during pregnancy, The initial therapeutic dose of LMWH varies with the speciﬁc without other high-risk conditions such as APA syndrome or anti-agent. Two agents have been approved by the U.S. Food and Drug thrombin deﬁciency, therapy should be continued in the postpartumAdministration (FDA) for the treatment of acute VTE: enoxaparin and period for 3 to 6 months. Because warfarin does not signiﬁcantlytinzaparin. Enoxaparin is given at a dose of 1 mg/kg administered accumulate in breast milk and does not induce an anticoagulant effectsubcutaneously twice daily (i.e., every 12 hours). For tinzaparin, the in the infant, it is not contraindicated in breastfeeding mothers.dose is 175 IU/kg administered subcutaneously once daily. A third Management of warfarin overdoses or hemorrhagic complicationsLMWH, dalteparin, has also been used off-label at doses of 100 U/kg is guided by the severity of the problem. For example, if patients areevery 12 hours. Barbour and colleagues evaluated whether the stan- found to have elevated INRs (>3.0) without bleeding, vitamin K candard therapeutic doses of dalteparin maintained peak therapeutic be given orally. However, if mild bleeding is present, vitamin K can belevels of anticoagulation during pregnancy; in 85% of patients, dosage administered subcutaneously.76 Normalization of the INR can occuradjustments were required to maintain peak anti-Xa activity between within 6 hours after a 5-mg oral or subcutaneous dose of vitamin K.0.5 and 1.0 IU/mL.72 Given these data and the fact that pregnancy Larger doses have a more rapid onset but render patients resistant topresents a period of rapidly changing volumes of distribution and re-anticoagulation with warfarin. In the setting of signiﬁcant hemor-ﬂuctuating concentrations of heparin-binding proteins, it appears rhage, fresh-frozen plasma will replenish clotting factors and can beprudent to monitor anti-Xa activity during therapeutic treatment with used with subcutaneous vitamin K to reverse the effects of warfarin.LMWH in pregnant patients and to adjust the doses to maintain atherapeutic level (i.e., anti-factor Xa levels of 0.6 to 1.2 U/mL) 4 hoursafter an injection. Complex Presentations As with unfractionated heparin, treatment should continue for 20weeks, and then prophylactic dosages should be given (e.g., enoxapa- Type 2 Heparin-Induced Thrombocytopeniarin, 40 mg SQ every 12 hr; dalteparin, 5000 U SQ once daily). There is In patients with a history or new presentation of HIT-2, heparin andcontroversy as to whether prophylactic LMWH warrants surveillance LMWH are contraindicated. Fondaparinux presents an excellent alter-with anti-factor Xa levels. If surveillance is undertaken, the goal should native. It is a synthetic heparin pentasaccharide that complexes withbe an anti-factor Xa level between 0.1 to 0.2 U/mL 4 hours after an the antithrombin binding site for heparin to permit the selective inac-injection. tivation of factor Xa but not thrombin. Excretion is renal, and the Patients with antithrombin deﬁciency, and patients who are homo- drug has a 15-hour half-life after a once-daily subcutaneous injection.zygotes or compound heterozygotes for the FVL or PGM mutation and Buller and associates conducted a randomized, double-blind trial ofhave a prior VTE or affected ﬁrst-degree relative, require therapeutic fondaparinux, administered subcutaneously once a day at a dose ofanticoagulation throughout pregnancy. In our practice, pregnant 5.0 mg for patients weighing less than 50 kg, 7.5 mg for those weighingpatients with these highly thrombogenic thrombophilias who are 50 to 100 kg, and 10.0 mg for those weighing more than 100 kg, versus
864 CHAPTER 41 Thromboembolic Disease in Pregnancyenoxaparin among 2205 patients with acute symptomatic DVT.77 No levels immediately before the next dose. Dose adjustments should bedifferences in recurrent VTE were observed between the two groups. made to insure that trough levels remain between 0.5 and 1.2 U/mL.83Fondaparinux is considered a pregnancy class B agent by the FDA. It In addition, the use of low-dose aspirin was recommended for suchhas been used in a small number of pregnant patients without adverse patients.83 These patients should be extensively counseled about thesequelae, although it has been found to be present in umbilical-cord risks and beneﬁts of these different regimens to both their own healthplasma at concentrations approximately 10% of those in the maternal and that of their fetus.plasma.78 These levels are well below those required for effective anti-coagulation. However, fondaparinux use in pregnant women is bestlimited to those patients with no obvious therapeutic alternatives, such Thrombolytic Therapyas patients with HIT-2 or severe allergic reaction to heparin. Although the mortality rate for expeditiously diagnosed and treated uncomplicated PE is less than 2% to 7%,84 rates higher than 50% haveThromboprophylaxis in Pregnant Patients with been reported for patients who were hemodynamically unstable at theMechanical Heart Valves time of presentation.85 This has led to more aggressive use of throm-There remains considerable controversy concerning optimal manage- bolytic therapy in patients with massive PE. However, meta-analysisment of VTE in pregnant women with mechanical heart valves. These of 9 randomized, controlled trials comparing thrombolytic agentspatients are given warfarin when in the nonpregnant state. However, versus intravenous heparin in patients with PE found thrombolyticwarfarin is loosely bound to albumin, readily crosses the placenta, and therapy offered no statistically signiﬁcantly different effect on mortal-is associated with an increased rate of birth defects (OR, 3.86; CI, 1.86 ity (RR, 0.63; CI, 0.32 to 1.23) or on recurrence of PE (RR, 0.59; CI,to 8.00) with exposure between 8 and 12 weeks’ gestation.79 The classic 0.30 to 1.18) compared with heparin, but such therapy was associatedfetal warfarin syndrome includes nasal hypoplasia, stippled epiphysis, with a signiﬁcantly increased risk of major hemorrhage (RR, 1.76; CI,and characteristic central nervous system defects including agenesis 1.04 to 2.98).86of the corpus callosum, Dandy-Walker malformation, midline cere- Pregnancy poses special concerns for thrombolytic therapy givenbellar atrophy, and ventral midline dysplasia with optic atrophy. the risk of abruption and puerperal hemorrhage. Turrentine and col-Maternal warfarin therapy after 12 weeks’ gestation has been associ- leagues reviewed the outcomes of 172 pregnancies treated with throm-ated with fetal and placental hemorrhage which can occur throughout bolytic therapy and reported a maternal mortality rate of 1.2%, a fetalpregnancy. loss rate of 6%, and maternal hemorrhagic complications in 8%.87 For these reasons, the agent is usually avoided during pregnancy. Although the data are limited, the risk of hemorrhage in the postpar-However, it may be appropriate to use warfarin in pregnant patients tum period appears to be limited to those treated within 8 hours ofwho have a mechanical heart valve. Meta-analysis suggests that, when delivery.88 Therefore, given the lack of clear beneﬁt and potentiallywarfarin is used throughout pregnancy the cumulative risk of embry- unique risks in decompensated pregnant patients refractory to heparinopathy in 6.4% (CI, 4.6% to 8.9%), but the risk of valvular thrombosis therapy, surgical thrombolectomy may be the preferred option.is quite low (3.9%; CI, 2.9% to 5.9%).80 In contrast, a regimen consist-ing of unfractionated heparin from 6 to 12 weeks, followed by warfarinuntil 36 weeks and then by unfractionated heparin until delivery,appears to reduce fetal risks but is associated with a substantially Preventionincreased risk of valve thrombosis (9.2%; CI, 5.9% to 13.9%). Warfarin is best employed in pregnant patients with mechanical Nonpharmacologic Preventionheart valves when the dosage can be kept lower than 5 mg/day, because A Cochrane review of randomized, controlled trials indicated that usecohort studies suggest that this dose is associated with a lower rate of of graduated compression stockings in hospitalized patients with pro-fetal complications.81 If warfarin is used in this setting, the target INR longed medical immobilization, or application of such stockings fromshould be 2.5 to 3.5. Low-dose aspirin should be used as an adjunct to before surgery until discharge or restoration of full mobility, reducedwarfarin, based on a study of antithrombotic therapy in high-risk the occurrence of DVT from 27% to 13% (OR, 0.34; CI, 0.25 to 0.46).89patients with mechanical valves.82 Warfarin therapy should be stopped A cohort study suggested that use of graduated elastic compressionby 36 weeks, and unfractionated heparin should then be administered stockings also reduced the prevalence of VTE in puerperal patients,subcutaneously every 8 to 12 hours, with doses adjusted to keep the from 4.3% to 0.9%.90 In addition, graduated elastic compression stock-aPTT value at 2 × control or the anti-Xa heparin level at 0.35 to ings have been shown to increase femoral vein ﬂow velocity in late0.70 U/mL. pregnancy.91 No large clinical studies exist to guide use of LMWH in pregnant Meta-analysis in nonpregnant patients with high or moderate riskpatients with mechanical heart valves. However, the manufacturer suggests that intermittent pneumatic compression devices decrease theof enoxaparin (Lovenox, Aventis, Bridgewater, NJ) speciﬁcally relative risk of DVT by 62% compared with placebo, by 47% comparedrecommends against its use in this setting, based on a small number with graduated compression stockings, and by 48% compared withof reports to the FDA of valvular thrombosis in pregnant women so low-dose unfractionated heparin.92 Because graduated elastic com-treated. The Anticoagulation in Prosthetic Valves and Pregnancy Con- pression stockings and pneumatic compression stockings have nosensus Report Panel and Scientiﬁc Roundtable analyzed these reported hemorrhagic risk and have been shown to be an effective means ofcases of valvular thrombosis in pregnant patients receiving LMWH for DVT prophylaxis in surgical patients and possibly in pregnant patients,mechanical heart valve prostheses in 2002 and concluded that virtually they should be strongly considered for prophylaxis in high-risk preg-all such cases were associated with underdosing or inadequate moni- nant patients (e.g., obesity, thrombophilia, strong family history) whotoring.83 Further, the panel recommended enoxaparin therapy in such are admitted for labor and/or delivery or who require prolonged bedpatients in lieu of warfarin, beginning at a dose of 1 mg/kg SQ every rest, and also in all pregnant patients undergoing an elective or repeat12 hours. They also recommended weekly monitoring of peak anti- cesarean delivery without prior labor. Caution must be exercisedfactor Xa levels 4 hours after injection, as well as monitoring trough in their immediate preoperative use after labor, because DVT may
CHAPTER 41 Thromboembolic Disease in Pregnancy 865have already formed and could be theoretically dislodged by either followed by initiation of the proper diagnostic algorithm. Treatmentdevice. requires prompt initiation of unfractionated heparin or LMWH. Finally, in pregnancy, left-lateral decubitus positioning during the Prevention includes identiﬁcation of high-risk patients and boththird trimester may also reduce the risk of VTE. nonpharmacologic and pharmacologic interventions.Pharmacologic PreventionAs noted, among pregnant patients who have had a previous VTE, References 1. Chang J, Elam-Evans LD, Berg CJ, et al: Pregnancy-related mortality sur-recurrence risks are highly dependent on the presence of a thrombo- veillance—United States, 1991-1999. MMWR Surveill Summ 52:1-8,philia and the nature of the risk factors associated with the prior 2003.thrombus.26 Whereas VTE recurrences were not observed in women 2. Gherman RB, Goodwin TM, Leung B, et al: Incidence, clinical characteris-without detectable thrombophilias whose previous VTE was associated tics, and timing of objectively diagnosed venous thromboembolism duringwith a temporary risk factor, 16% of thrombophilic patients who did pregnancy. Obstet Gynecol 94:730-734, 1999.not receive thromboprophylaxis during pregnancy had a recurrent 3. Stein PD, Henry JW: Prevalence of acute pulmonary embolism amongVTE.26 However, even the former group of patients require postpar- patients in a general hospital and at autopsy. Chest 108:978-981, 1995. 4. Cushman M: Epidemiology and risk factors for venous thrombosis. Semintum thromboprophylaxis. In addition, it has been argued that throm- Hematol 44:62-69, 2007.bophilic patients who are without a personal history of VTE but who 5. Bremme KA: Haemostatic changes in pregnancy. Best Pract Res Clinhave an affected ﬁrst-degree relative should also receive thrombopro- Haematol 16:153-168, 2003.phylaxis during pregnancy and the postpartum period. Those with 6. Paidas MJ, Ku DH, Lee MJ, et al: Protein Z, protein S levels are lower inhighly thrombogenic thrombophilias (e.g., antithrombin deﬁciency, patients with thrombophilia and subsequent pregnancy complications.homozygosity or compound heterozygosity for the FVL or PGM J Thromb Haemost 3:497-501, 2005.mutation) should receive both therapeutic or subtherapeutic doses 7. de Boer K, ten Cate JW, Sturk A, et al: Enhanced thrombin generation inof LMWH throughout pregnancy and postpartum anticoagulation. normal and hypertensive pregnancies. Am J Obstet Gynecol 160:95-100,However, such therapy does not appear to be justiﬁed during the 1989.antepartum period in patients with less thrombogenic thrombophilias 8. Cockett FB, Thomas ML, Negus D: Iliac vein compression: Its relation to iliofemoral thrombosis and the post-thrombotic syndrome. BMJ 2:14-19,(e.g., heterozygosity for FVL or PGM, protein C deﬁciency, protein 1967.S deﬁciency) who are without a personal or strong family history of 9. Edouard DA, Pannier BM, London GM, et al: Venous and arterial behaviorVTE. However, these patients should receive postpartum anticoagula- during normal pregnancy. Am J Physiol 274:H1605-H1612, 1998.tion if they require a cesarean delivery, to reduce the risk of a fatal 10. Carbillon L, Uzan M, Uzan S: Pregnancy, vascular tone, and maternalPE. hemodynamics: A crucial adaptation. Obstet Gynecol Surv 55:574-581, A very limited number of studies have assessed the value of peri- 2000.operative thromboprophylaxis in cesarean delivery.93,94 However, peri- 11. Macklon NS, Greer IA: Venous thromboembolic disease in obstetrics andoperative thromboprophylaxis with low-dose unfractionated heparin gynaecology: The Scottish experience. Scott Med J 41:83-86, 1996.may be appropriate for patients undergoing cesarean delivery who 12. Chisaka H, Utsunomiya H, Okamura K, Yaegashi N: Pulmonary thrombo-have a history of VTE and known highly thrombogenic thrombophilia, embolism following gynecologic surgery and cesarean section. Int J Gyn- aecol Obstet 84:47-53, 2004.as well as those with mechanical heart valve prostheses. 13. Clark CA, Spitzer KA, Crowther MA, et al: Incidence of postpartum throm- bosis and preterm delivery in women with antiphospholipid antibodies and recurrent pregnancy loss. J Rheumatol 34:996-996, 2007.Inferior Vena Cava Filters 14. Branch DW, Silver RM, Blackwell JL, et al: Outcome of treated pregnanciesInferior vena cava ﬁlters are designed for use in patients in whom in women with antiphospholipid syndrome: An update of the Utah experi-anticoagulation is absolutely contraindicated, such as those with a ence. Obstet Gynecol 80:614-620, 1992.hemorrhagic stroke, recent or current hemorrhage, or recent surgery. 15. Ogunyemi D, Cuellar F, Ku W, Arkel Y: Association between inheritedThis intervention appears to be appropriate in patients who have thrombophilias, antiphospholipid antibodies, and lipoprotein A levels andrecurrent PE despite adequate anticoagulation and in those in whom venous thromboembolism in pregnancy. Am J Perinatol 20:17-24, 2003.a PE would probably be lethal (e.g., patients with pulmonary hyperten- 16. Gerhardt A, Scharf RE, Beckmann MW, et al: Prothrombin and factor V mutations in women with a history of thrombosis during pregnancy andsion). Pregnant patients with a history of HIT-2 were traditional can- the puerperium. N Engl J Med 342:374-380, 2000.didates for inferior vena caval ﬁlters, but the use of fondaparinux 17. Zotz RB, Gerhardt A, Scharf RE: Inherited thrombophilia and gestationalduring pregnancy and the direct thrombin inhibitors (e.g., Lepirudin) venous thromboembolism. Best Pract Res Clin Haematol 16:243-259,in the postpartum period has rendered this indication less absolute.95 2003.Although the use of ﬁlters is generally discouraged in younger patients, 18. Friederich PW, Sanson BJ, Simioni P, et al: Frequency of pregnancy-relatedretrievable ﬁlters have been employed successfully in pregnant women venous thromboembolism in anticoagulant factor-deﬁcient women:and may prove ideal in this setting.96 implications for prophylaxis. Ann Intern Med 125:955-960, 1996. 19. Franco RF, Reitsma PH: Genetic risk factors of venous thrombosis. Hum Genet 109:369-384, 2001. 20. Carraro P, European Communities Confederation of Clinical ChemistrySummary and Laboratory Medicine, Working Group on Guidelines for Investigation of Disease: Guidelines for the laboratory investigation of inherited throm-Acute PE remains a leading cause of maternal morbidity and mortality. bophilias: Recommendations for the ﬁrst level clinical laboratories. ClinPeriods of maximal risk include the immediate puerperal perioperative Chem Lab Med 41:382-391, 2003.period. Additional risk factors include prior VTE, positive family 21. Goodwin AJ, Rosendaal FR, Kottke-Marchant K, Bovill EG: A review of thehistory, obesity, thrombophilias, infection, trauma, and immobiliza- technical, diagnostic, and epidemiologic considerations for protein Stion. The prompt diagnosis of VTE requires assessment of clinical risk assays. Arch Pathol Lab Med 126:1349-1366, 2002.
866 CHAPTER 41 Thromboembolic Disease in Pregnancy22. Conard J, Horellou MH, Van Dreden P, et al: Thrombosis and pregnancy 46. Kasper W, Meinertz T, Kersting F, et al: Echocardiography in assessing in congenital deﬁciencies in AT III, protein C or protein S: Study of 78 acute pulmonary hypertension due to pulmonary embolism. Am J Cardiol women. Thromb Haemost 63:319-320, 1990. 45:567-572, 1980.23. Brill-Edwards P, Ginsberg JS, Gent M, et al; Recurrence of Clot in 47. Pruszczyk P, Torbicki A, Pacho R, et al: Noninvasive diagnosis of suspected This Pregnancy Study Group: Safety of withholding heparin in pregnant severe pulmonary embolism: Transesophageal echocardiography versus women with a history of venous thromboembolism. N Engl J Med spiral CT. Chest 112:722-728, 1997. 343:1439-1444. 48. Adachi T, Umezaki I, Okano H, et al: Placenta previa totalis complicated24. Ginsberg JS: Management of venous thromboembolism. N Engl J Med with pulmonary embolism during cesarean section: A case report. Semin 335:1816-1828, 1996. Thromb Hemost 31:321-326, 2005.25. Hirsh J, Hoak J: Management of deep vein thrombosis and pulmonary 49. Winer-Muram HT, Boone JM, Brown HL, et al: Pulmonary embolism in embolism: A statement for healthcare professionals from the Council pregnant patients: Fetal radiation dose with helical CT. Radiology 224:487- on Thrombosis (in consultation with the Council on Cardiovascular 492, 2002. Radiology), American Heart Association. Circulation 93:2212-2245, 50. The PIOPED Investigators: Value of the ventilation/perfusion scan in acute 1996. pulmonary embolism: Results of the Prospective Investigation Of Pulmo-26. Wells PS, Anderson DR, Rodger M, et al: Evaluation of D-dimer in the nary Embolism Diagnosis (PIOPED). JAMA 263:2753-2759, 1990. diagnosis of suspected deep-vein thrombosis. N Engl J Med 349:1227-1235, 51. Chan WS, Ray JG, Murray S, et al: Suspected pulmonary embolism 2003. in pregnancy: Clinical presentation, results of lung scanning, and subse-27. Kassai B, Boissel JP, Cucherat M, et al: A systematic review of the accuracy quent maternal and pediatric outcomes. Arch Intern Med 162:1170-1175, of ultrasound in the diagnosis of deep venous thrombosis in asymptomatic 2002. patients. Thromb Haemost 91:655-666, 2004. 52. Cross JJ, Kemp PM, Walsh CG, et al: A randomized trial of spiral CT and28. Gottlieb RH, Widjaja J, Tian L, et al: Calf sonography for detecting deep ventilation perfusion scintigraphy for the diagnosis of pulmonary embo- venous thrombosis in symptomatic patients: Experience and review of the lism. Clin Radiol 53:177-182, 1998. literature. J Clin Ultrasound 27:415-420, 1999. 53. Moores LK, Jackson WL Jr, Shorr AF, Jackson JL: Meta-analysis: Outcomes29. Andrews EJ Jr, Fleischer AC: Sonography for deep venous thrombosis: in patients with suspected pulmonary embolism managed with computed Current and future applications. Ultrasound Q 21:213-225, 2005. tomographic pulmonary angiography. Ann Intern Med 141:866-874,30. Evans AJ, Sostman HD, Witty LA, et al: Detection of deep venous throm- 2004. bosis: Prospective comparison of MR imaging and sonography. J Magn 54. Remy-Jardin M, Remy J, Baghaie F, et al: Clinical value of thin collimation Reson Imaging 6:44-51, 1996. in the diagnostic workup of pulmonary embolism. AJR Am J Roentgenol31. Tapson VF, Carroll BA, Davidson BL, et al: The diagnostic approach to 175:407-411, 2000. acute venous thromboembolism: Clinical practice guideline. American 55. Difﬁn D, Leyendecker JR, Johnson SP, et al: Effect of anatomic dis- Thoracic Society. Am J Respir Crit Care Med 160:1043-1066, 1999. tribution of pulmonary emboli on interobserver agreement in the inter-32. Hirsh J, Lee AY: How we diagnose and treat deep vein thrombosis. Blood pretation of pulmonary angiography. AJR Am J Roentgenol 171:1085-1089, 99:3102-3110, 2002. 1998.33. Wells PS, Brill-Edwards P, Stevens P, et al: D-dimer testing for deep venous 56. Stein PD, Henry JW, Gottschalk A: Reassessment of pulmonary angiogra- thrombosis: A metaanalysis. Circulation 91:2184-2187, 1995. phy for the diagnosis of pulmonary embolism: Relation of interpreter34. Morse M: Establishing a normal range for D-dimer levels through preg- agreement to the order of the involved pulmonary arterial branch. Radiol- nancy to aid in the diagnosis of pulmonary embolism and deep vein throm- ogy 210:689-691, 1999. bosis. J Thromb Haemost 2:1202-1204, 2004. 57. Meaney JF, Weg JG, Chenevert TL, et al: Diagnosis of pulmonary embolism35. Chan WS, Chunilal SD, Lee AY, et al: Diagnosis of deep vein thrombosis with magnetic resonance angiography. N Engl J Med 336:1422-1427, during pregnancy: A pilot study evaluating the role of d-dimer and com- 1997. pression leg ultrasound during pregnancy. Blood 100:275A, 2002. 58. Oudkerk M, van Beek EJ, Wielopolski P, et al: Comparison of contrast-36. Epiney M, Boehlen F, Boulvain M, et al: D-dimer levels during delivery and enhanced magnetic resonance angiography and conventional pulmonary the postpartum. J Thromb Haemost 3:268-271, 2005. angiography for the diagnosis of pulmonary embolism: A prospective37. Koh SC, Pua HL, Tay DH, Ratnam SS: The effects of gynaecological study. Lancet 359:1643-1647, 2002. surgery on coagulation activation, ﬁbrinolysis and ﬁbrinolytic inhibitor in 59. Mills SR, Jackson DC, Older RA, et al: The incidence, etiologies, and avoid- patients with and without ketorolac infusion. Thromb Res 79:501-514, ance of complications of pulmonary angiography in a large series. Radiol- 1995. ogy 136:295-299, 1980.38. Prandoni P, Cogo A, Bernardi E, et al: A simple ultrasound approach for 60. Dalen JE, Brooks HL, Johnson LW, et al: Pulmonary angiography in acute detection of recurrent proximal-vein thrombosis. Circulation 88:1730- pulmonary embolism: Indications, techniques, and results in 367 patients. 1735, 1993. Am Heart J 81:175-185, 1971.39. Stein PD, Terrin ML, Hales CA, et al: Clinical, laboratory, roentgeno- 61. Brent RL: The effect of embryonic and fetal exposure to x-ray, microwaves, graphic, and electrocardiographic ﬁndings in patients with acute pulmo- and ultrasound: Counseling the pregnant and nonpregnant patient about nary embolism and no pre-existing cardiac or pulmonary disease. Chest these risks. Semin Oncol 16:347-368, 1989. 100:598-603, 1991. 62. Webb JA, Thomsen HS, Morcos SK; Members of Contrast Media Safety40. Fedullo PF, Tapson VF: Clinical practice: The evaluation of suspected Committee of European Society of Urogenital Radiology (ESUR): The use pulmonary embolism. N Engl J Med 349:1247-1256, 2003. of iodinated and gadolinium contrast media during pregnancy and41. Scarsbrook AF, Gleeson FV: Investigating suspected pulmonary embolism lactation. Eur Radiol 15:1234-1240, 2005. in pregnancy. BMJ 334:418-419, 2007. 63. Kearon C: Diagnosis of pulmonary embolism. CMAJ 168:183-194, 2003.42. The Urokinase Pulmonary Embolism Trial: A national cooperative study. 64. Stein PD, Hull RD, Saltzman HA, Pineo G: Strategy for diagnosis of patients Circulation 47(Suppl 2):1-108, 1973. with suspected pulmonary embolism. Chest 103:1553-1559, 1993.43. Costantini M, Bossone E, Renna R, et al: Electrocardiographic features 65. Kearon C, Ginsberg JS, Douketis J, et al; Canadian Pulmonary Embolism in critical pulmonary embolism: Results from baseline and continuous Diagnosis Study (CANPEDS) Group: An evaluation of D-dimer in the electrocardiographic monitoring. Ital Heart J 5:214-216. diagnosis of pulmonary embolism: A randomized trial. Ann Intern Med44. Green RM, Meyer TJ, Dunn M, Glassroth J: Pulmonary embolism in 144:812-821, 2006. younger adults. Chest 101:1507-1511, 1992. 66. Stein PD, Hull RD, Patel KC, et al: D-dimer for the exclusion of acute45. Come PC: Echocardiographic evaluation of pulmonary embolism and its venous thrombosis and pulmonary embolism: A systematic review. Ann response to therapeutic interventions. Chest 101:151S-162S, 1992. Intern Med 140:589-602, 2004.