Ultrasonographic diagnosis of symptomatic deep venous thrombosis in pregnancy Joseph F. Polak, MD, and Diane L. Wilkinson, MD Boston, Massachusetts Color flow imaging combined with compression ultrasonography were used to prospectively evaluate symptomatic patients first seen with suspected deep vein thrombosis during pregnancy. The incidence of symptomatic deep vein thrombosis (7 of 10,000 deliveries) shown prospectively by ultrasonography was similar to that of historic controls who had a diagnosis confirmed by venography (11 of 26,191 deliveries). A preferential pattern of proximal left-sided iliac or femoral vein involvement was common to both cohorts, being seen in five of the seven cases diagnosed by ultrasonography and eight of the eleven cases diagnosed by venography. We conclude that ultrasonography, conSisting of compression ultrasonography and color flow imaging, can replace venography in the diagnosis and evaluation of the extent of symptomatic deep vein thrombosis during pregnancy. (AM J OBSTET GVNECOL 1991 ;165:625-9.)
Key words: Venous thrombosis, ultrasonography, pregnancy Pregnancy is a well-recognized risk factor in the development of deep venous thrombosis in women younger than 40 years of age.! This diagnosis has until recently been made on clinical grounds, l. 2 by venography,3 or with the use of impedance plethysmography"'s Compression ultrasonography has been shown to have an accuracy >95% for the diagnosis of venous thrombi located within the femoral and popliteal venous segments. 5 . 9 Although it might be possible to infer that the same accuracy is attained in the pregnant patient, this issue has only been addressed in part in a report of three cases.1O This study proposes to evaluate the utility of color flow imaging and compression ultrasonography in the diagnosis and evaluation of the extent of symptomatic deep vein thrombosis in the antepartum patient.
Material and methods Prospective cohort. Antepartum patients first seen with clinical suspected symptomatic lower extremity venous thrombosis were prospectively studied between March 1988 and March 1989. During this time interval, 10,000 deliveries were performed at our institution. There were 28 patients who ranged in age from 17 to 40 years (mean, 26.8 years ± 6.1 SD). When first seen, patients were on average 26.4 ± 9.5 weeks into their pregnancy for a range of 6 to 39 weeks. All complained
From the Department of Radiology, Harvard Medical School, Brigham and Women's Hospital. Received for publication September 6, 1990; revised january 14, 1991; accepted March 1, 1991. Reprint requests: joseph F. Polak, MD, Department of Radiology, Brigham and Women's Hospital, 75 Francis St., Boston, MA 02115. 6/1 /29183
of leg swelling, eight had additional complaints of calf pain, and eight localized the pain to the thigh. Retrospective cohort. The medical records and venograms of all antepartum patients identified retrospectively as having had venography performed between August 1984 and April 1987 were reviewed according to institutional guidelines. Twenty-two such cases were identified. The number of deliveries during this time period was 26,191. The average age of this group was 24.4 years ± 7.0 SD (range, 18 to 41 years). When first seen, these patients were on average 25 ± 8.1 weeks into their pregnancy for a range of 8 to 38 weeks. Nonpregnant cohort. Between October 1988 and March 1989, 51 nonpregnant patients had both ultrasonography and venography performed because of the clinical suspicion of deep vein thrombosis of the lower extremity. There were 23 males and 28 females with an average age of 54.8 years ± 16.8 SD (range, 18 to 89 years). Forty-three patients were referred because of symptoms, whereas eight were undergoing postoperative surveillance for the presence of deep vein thrombosis. There were 39 inpatients and 12 outpatients. Techniques. Ultrasonography was performed with the aid of a 5 MHz linear array transducer (Acuson, Computed Sonography, Mountain View, Calif.). The imaging protocol consisted of using distention of the common femoral vein in response to the Val salva maneuver to exclude obstructing thrombus in the iliac veins.!! The remainder of the femoral popliteal venous system was then surveyed with color Doppler flow imaging with flow augmentation in the femoral and popliteal vein to exclude nonobstructing thrombosis. Compression ultrasonography was also performed
625
626
Polak and Wilkinson
Seplember 1991
Am J ObSlCl Gynccol
Fig. lAo This transverse view of the left common femoral vein shows a distended vein (arrows) medial to the artery (arrowheads).
Fig. lB. Progressive compression causes deformation of the artery (curved arrow) before the vein. This loss of compressibility in spite of the lack of echogenic signals within the vein is diagnostic of venous thrombosis.
along the course of the femoral and popliteal veins with the transducer held transverse to the vein. All patients were examined in the supine position. The popliteal vein was imaged with the extremity in slight external rotation or froglike position. Deep vein thrombosis was diagnosed when a venous segment was noncompressible and did not show any flow signals. If the patient manifested localized symptoms a survey was also performed at the site of symptoms, whether it be in the calf or in the thigh.'2 Venography. Venography was performed with a modification of the approach described by Rabinov and Paulin'" with a 23 or 25 gauge needle used to cannulate
a dorsal pedal vein and dilute contrast material injected under fluoroscopic guidance. Data analysis. Comparison of the distribution of venous thrombi was made using X2 statistics. Proximal venous thrombosis refers to either iliac, common femoral, or superficial femoral vein involvement with sparing of the popliteal veins. Distal venous thrombosis is limited to the calf and popliteal veins with sparing of either the iliac, common femoral, or superficial femoral veins. Sensitivity and specificity were determined according to recommendations made by McNeil et al." The sensitivity and specificity of the ultrasonographic imaging
Venous thrombosis in pregnancy
Volullle lG:,
627
~umb('r ~
100
• ~
80
'"OJ '"01 U
Pregnant Non Pregnant
60
... 0
~
40
20
0
Right Side
Left Side
Fig. 2. The figure depicts the affected side of venous thrombosis in the patients studied. Comparison of the side involved by deep vein thrombosis in pregnant and nonpregnant patients shows the preferential left-sided involvement during pregnancy.
protocol was evaluated by comparison to the phlebographic results in the non pregnant patients because none of the pregnant patients with ultrasonographic examinations underwent additional phlebography. Results
Pregnant subjects. In the retrospective cohort, there were 22 patients in whom venography was performed out of 26,191 pregnancies at risk. In 11 of these, the diagnosis of deep venous thrombosis was confirmed. Thrombosis of the lower popliteal veins was noted in three, whereas proximal thrombus sparing the popliteal vein but involving the iliofemoral system was noted in the other eight. The left side was involved in all 11 cases. Venous thrombosis occurred during the first (n = 2), second (n = 5), and third (n = 4) trimesters. In the prospective cohort diagnosed by ultrasonography, 28 subjects were suspected of having deep venous thrombosis. The diagnosis was made in seven instances by noting loss of compressibility and the absence of flow within the involved segment. Proximal venous thrombosis involving the iliac and femoral vein but sparing the popliteal vein was diagnosed in five of these (Figs. lA and IB). The left side was involved in five of the seven cases. Deep vein thrombosis occurred during the first (n = 2), second (n = 1), and third (n = 4) trimesters. All patients diagnosed by ultrasonography or venography were treated with heparin and went on to be delivered of infants. No Cases of clinically evident pulmonary emboli were reported. Nonpregnant subjects. A total of 45 of 51 venograms were judged to be diagnostic with three being technically limited and three nondiagnostic. There were 13 instances of negative venograms; all
13 had a normal noninvasive ultrasonographic examination. Three of seven cases of below-knee thrombosis were prospectively detected. The 25 instances of femoropopliteal venous thrombosis were diagnosed prospectively by ultrasonography. There was a slight pref~ erentialleft lower extremity involvement (14/25). Distal thrombosis involving the popliteal and femoral veins but sparing the iliac veins was seen in 22 of 25 cases. Proximal thrombosis limited to the iliofemoral veins with sparing of the popliteal veins was seen in three patients with pelvic malignancies. Pregnant vs nonpregnant subjects. Preferential leftsided involvement (Fig. 2) was more common in the pregnant population (16/18 vs 18/32; X" = 4.24; P < 0.05). Proximal iliofemoral thrombosis was more common in the pregnant population (Fig. 3) as compared with the nonpregnant population (13/18 vs 3/32; X" = 18.1; P < 0.001). Comment
This study confirms that the incidence of symptomatic deep venous thrombosis during pregnancy is low and similar to that cited in previous reports (i.e., between 0.4 to 0.7/1000 deliveries).' I Among subjects younger than 40 years of age, pregnancy is a major risk factor for venous thrombosis and pulmonary embolism. I Once recognized, venous thrombosis is treated with judicious anticoagulation. te, Iii The use of heparin is favored, in spite of the logistical difficulties of its administration, because of the possible teratogenic effects of sodium warfarin (Coumadin). A previous report of three cases suggested that venous ultrasonography could be used to make the diagnosis of deep vein thrombosis during pregnancy. 10
628
Polak and Wilkinson All!
ScplclI!ber 19'11 CVlleml
J ObslCI
100
80
.. til
•
Pregnant
~
Non Pregnant
60
til
...
Key words: Venous thrombosis, ultrasonography, pregnancy Pregnancy is a well-recognized risk factor in the development of deep venous thrombosis in women younger than 40 years of age.! This diagnosis has until recently been made on clinical grounds, l. 2 by venography,3 or with the use of impedance plethysmography"'s Compression ultrasonography has been shown to have an accuracy >95% for the diagnosis of venous thrombi located within the femoral and popliteal venous segments. 5 . 9 Although it might be possible to infer that the same accuracy is attained in the pregnant patient, this issue has only been addressed in part in a report of three cases.1O This study proposes to evaluate the utility of color flow imaging and compression ultrasonography in the diagnosis and evaluation of the extent of symptomatic deep vein thrombosis in the antepartum patient.
Material and methods Prospective cohort. Antepartum patients first seen with clinical suspected symptomatic lower extremity venous thrombosis were prospectively studied between March 1988 and March 1989. During this time interval, 10,000 deliveries were performed at our institution. There were 28 patients who ranged in age from 17 to 40 years (mean, 26.8 years ± 6.1 SD). When first seen, patients were on average 26.4 ± 9.5 weeks into their pregnancy for a range of 6 to 39 weeks. All complained
From the Department of Radiology, Harvard Medical School, Brigham and Women's Hospital. Received for publication September 6, 1990; revised january 14, 1991; accepted March 1, 1991. Reprint requests: joseph F. Polak, MD, Department of Radiology, Brigham and Women's Hospital, 75 Francis St., Boston, MA 02115. 6/1 /29183
of leg swelling, eight had additional complaints of calf pain, and eight localized the pain to the thigh. Retrospective cohort. The medical records and venograms of all antepartum patients identified retrospectively as having had venography performed between August 1984 and April 1987 were reviewed according to institutional guidelines. Twenty-two such cases were identified. The number of deliveries during this time period was 26,191. The average age of this group was 24.4 years ± 7.0 SD (range, 18 to 41 years). When first seen, these patients were on average 25 ± 8.1 weeks into their pregnancy for a range of 8 to 38 weeks. Nonpregnant cohort. Between October 1988 and March 1989, 51 nonpregnant patients had both ultrasonography and venography performed because of the clinical suspicion of deep vein thrombosis of the lower extremity. There were 23 males and 28 females with an average age of 54.8 years ± 16.8 SD (range, 18 to 89 years). Forty-three patients were referred because of symptoms, whereas eight were undergoing postoperative surveillance for the presence of deep vein thrombosis. There were 39 inpatients and 12 outpatients. Techniques. Ultrasonography was performed with the aid of a 5 MHz linear array transducer (Acuson, Computed Sonography, Mountain View, Calif.). The imaging protocol consisted of using distention of the common femoral vein in response to the Val salva maneuver to exclude obstructing thrombus in the iliac veins.!! The remainder of the femoral popliteal venous system was then surveyed with color Doppler flow imaging with flow augmentation in the femoral and popliteal vein to exclude nonobstructing thrombosis. Compression ultrasonography was also performed
625
626
Polak and Wilkinson
Seplember 1991
Am J ObSlCl Gynccol
Fig. lAo This transverse view of the left common femoral vein shows a distended vein (arrows) medial to the artery (arrowheads).
Fig. lB. Progressive compression causes deformation of the artery (curved arrow) before the vein. This loss of compressibility in spite of the lack of echogenic signals within the vein is diagnostic of venous thrombosis.
along the course of the femoral and popliteal veins with the transducer held transverse to the vein. All patients were examined in the supine position. The popliteal vein was imaged with the extremity in slight external rotation or froglike position. Deep vein thrombosis was diagnosed when a venous segment was noncompressible and did not show any flow signals. If the patient manifested localized symptoms a survey was also performed at the site of symptoms, whether it be in the calf or in the thigh.'2 Venography. Venography was performed with a modification of the approach described by Rabinov and Paulin'" with a 23 or 25 gauge needle used to cannulate
a dorsal pedal vein and dilute contrast material injected under fluoroscopic guidance. Data analysis. Comparison of the distribution of venous thrombi was made using X2 statistics. Proximal venous thrombosis refers to either iliac, common femoral, or superficial femoral vein involvement with sparing of the popliteal veins. Distal venous thrombosis is limited to the calf and popliteal veins with sparing of either the iliac, common femoral, or superficial femoral veins. Sensitivity and specificity were determined according to recommendations made by McNeil et al." The sensitivity and specificity of the ultrasonographic imaging
Venous thrombosis in pregnancy
Volullle lG:,
627
~umb('r ~
100
• ~
80
'"OJ '"01 U
Pregnant Non Pregnant
60
... 0
~
40
20
0
Right Side
Left Side
Fig. 2. The figure depicts the affected side of venous thrombosis in the patients studied. Comparison of the side involved by deep vein thrombosis in pregnant and nonpregnant patients shows the preferential left-sided involvement during pregnancy.
protocol was evaluated by comparison to the phlebographic results in the non pregnant patients because none of the pregnant patients with ultrasonographic examinations underwent additional phlebography. Results
Pregnant subjects. In the retrospective cohort, there were 22 patients in whom venography was performed out of 26,191 pregnancies at risk. In 11 of these, the diagnosis of deep venous thrombosis was confirmed. Thrombosis of the lower popliteal veins was noted in three, whereas proximal thrombus sparing the popliteal vein but involving the iliofemoral system was noted in the other eight. The left side was involved in all 11 cases. Venous thrombosis occurred during the first (n = 2), second (n = 5), and third (n = 4) trimesters. In the prospective cohort diagnosed by ultrasonography, 28 subjects were suspected of having deep venous thrombosis. The diagnosis was made in seven instances by noting loss of compressibility and the absence of flow within the involved segment. Proximal venous thrombosis involving the iliac and femoral vein but sparing the popliteal vein was diagnosed in five of these (Figs. lA and IB). The left side was involved in five of the seven cases. Deep vein thrombosis occurred during the first (n = 2), second (n = 1), and third (n = 4) trimesters. All patients diagnosed by ultrasonography or venography were treated with heparin and went on to be delivered of infants. No Cases of clinically evident pulmonary emboli were reported. Nonpregnant subjects. A total of 45 of 51 venograms were judged to be diagnostic with three being technically limited and three nondiagnostic. There were 13 instances of negative venograms; all
13 had a normal noninvasive ultrasonographic examination. Three of seven cases of below-knee thrombosis were prospectively detected. The 25 instances of femoropopliteal venous thrombosis were diagnosed prospectively by ultrasonography. There was a slight pref~ erentialleft lower extremity involvement (14/25). Distal thrombosis involving the popliteal and femoral veins but sparing the iliac veins was seen in 22 of 25 cases. Proximal thrombosis limited to the iliofemoral veins with sparing of the popliteal veins was seen in three patients with pelvic malignancies. Pregnant vs nonpregnant subjects. Preferential leftsided involvement (Fig. 2) was more common in the pregnant population (16/18 vs 18/32; X" = 4.24; P < 0.05). Proximal iliofemoral thrombosis was more common in the pregnant population (Fig. 3) as compared with the nonpregnant population (13/18 vs 3/32; X" = 18.1; P < 0.001). Comment
This study confirms that the incidence of symptomatic deep venous thrombosis during pregnancy is low and similar to that cited in previous reports (i.e., between 0.4 to 0.7/1000 deliveries).' I Among subjects younger than 40 years of age, pregnancy is a major risk factor for venous thrombosis and pulmonary embolism. I Once recognized, venous thrombosis is treated with judicious anticoagulation. te, Iii The use of heparin is favored, in spite of the logistical difficulties of its administration, because of the possible teratogenic effects of sodium warfarin (Coumadin). A previous report of three cases suggested that venous ultrasonography could be used to make the diagnosis of deep vein thrombosis during pregnancy. 10
628
Polak and Wilkinson All!
ScplclI!ber 19'11 CVlleml
J ObslCI
100
80
.. til
•
Pregnant
~
Non Pregnant
60
til
...
