- Email: [email protected]
Risk factor scoring for predicting venous thromboembolism in obstetric patients
Risk factor scoring for predicting venous thromboembolism in obstetric patients Nofit Weiss, MD, and Peter S. Bernstein, MD, MPH Bronx, New York OBJECTIVE: The aim of this study was to develop a risk factor scoring system for the prediction of venous thromboembolism in obstetric patients. STUDY DESIGN: We conducted a retrospective case-control study of all pregnant or postpartum women admitted to the Jack D. Weiler Hospital from 1987 through 1998 with a discharge diagnosis indicating thromboembolism. For each study subject the three women who were delivered immediately before that index patient were selected for the control group. Data collected included the following: history of thrombosis, age, body mass index, previous abdominal surgical procedures, presence of systemic diseases, and blood type. Each patient was assigned a score that was based on the risk factors identified. RESULTS: We identified 21 patients who had sustained thromboembolic events during pregnancy or 6 weeks post partum. Nineteen of the thromboembolic events (90%) were diagnosed during pregnancy, and these cases were distributed throughout gestation—8 (42%) in the first trimester, 2 (10%) in the second trimester, and 9 (48%) in the third trimester. Six (28%) of these patients had pulmonary embolisms. Two cases of postpartum thromboembolic events were documented. Both were cases of pulmonary embolism. There was 1 maternal death. Patients with a score >2 were at significantly increased risk for having a thromboembolism, with an odds ratio of 4.8 (P < .05). The sensitivity, specificity, positive predictive value, and negative predictive value of this cutoff point were 21%, 95%, 57%, and 78%, respectively. CONCLUSION: Obstetric patients with high risk factor scores were at increased risk for thromboembolism. (Am J Obstet Gynecol 2000;182:1073-5.)
Key words: Deep vein thrombosis, pregnancy, pulmonary embolism, risk factor scoring, thromboembolism
Thromboembolic events are a major potentially preventable cause of maternal morbidity and mortality, with a quoted incidence of pulmonary embolism during pregnancy between 0.09 and 0.7 per 1000 live births.1-3 Women have a 5-fold increased risk of venous thrombosis during pregnancy,4 and this frequency increases with any additional risk factors.1 Pulmonary embolism will develop in as many as 24% of patients with untreated deep vein thrombosis, resulting in a mortality rate of approximately 13%.5 Among patients with deep vein thromboses who receive proper anticoagulation, pulmonary embolism occurs in 4.5% and the mortality rate is reduced to <1%.6 To date there have been no large randomized trials in the obstetric literature examining the efficacy of thromboembolic prophylaxis during pregnancy and the postpartum period. Estimates of occurrence and recurrence
From the Department of Obstetrics and Gynecology and Women’s Health, Albert Einstein College of Medicine, and Montefiore Medical Center. Reprint requests: Peter S. Bernstein, MD, Jack D. Weiler Hospital of the Albert Einstein College of Medicine, 1825 Eastchester Rd, Room 703, Bronx, NY 10461-2373. Copyright © 2000 by Mosby, Inc. 0002-9378/2000 $12.00 + 0 6/1/105441 doi:10.1067/mob.2000.105441
in pregnancy have not been stratified by risk factors and have been based on retrospective studies. Many of the recommendations from consensus panels regarding prophylactic strategies in pregnancy have been based on trials with nonpregnant patients. A recent survey of maternal-fetal medicine specialists revealed a lack of consensus regarding indications for and use of peripartum thromboembolic prophylaxis.7 This confusion regarding how best to prevent thromboembolism among pregnant and postpartum women reflects the lack of good data to guide management. The development of thromboembolism risk factor scoring systems has been suggested in the surgical and gynecologic literature.8-12 Different indexes have been reported to predict the risk for development of postoperative thromboembolic events in different surgical groups. The application of such indexes allows targeting of appropriate prophylaxis. For example, Caprini et al8 proposed a risk assessment rating system that assigned a predetermined score to each risk factor. On the basis of the overall scores specific modalities for thromboembolic prophylaxis were recommended. Clayton et al9 in the surgical literature and ClarkePearson et al13 in the gynecologic literature have proposed similar prognostic indexes. Although there are 1073
1074 Weiss and Bernstein
May 2000 Am J Obstet Gynecol
Table I. Obstetric patients with thromboembolism and their assigned scores Patient
Thromboembolic event
Gestational age (wk)
Score
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Pulmonary embolism Deep venous thrombosis Deep venous thrombosis Pulmonary embolism Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Pulmonary embolism Pulmonary embolism Deep venous thrombosis Pulmonary embolism Pulmonary embolism
28 37 28 30 34 12 37 24 12 08 05 11 38 10 09 34 22 30 08 1 d post partum* 5 wk post partum*
4 1 2 0 0 1 1 2 1 1 2 0 0 5 1 3 1 5 1 2 3
Risk factors Previous thrombosis Non-O blood type Age ≥35 y, previous abdominal surgery Body mass index >29.0 kg/m2 Previous abdominal surgery Non-O blood type, body mass index >29.0 kg/m2 Previous abdominal surgery Non-O blood type Non-O blood type, body mass index >29.0 kg/m2 Previous thrombosis, non-O blood type Non-O blood type Non-O blood type, age ≥35 y, sickle cell disease Previous abdominal surgery Previous thrombosis, non-O blood type Non-O blood type Cesarean delivery, body mass index >29.0 kg/m2 Age ≥35 y, non-O blood type, body mass index >29.0 kg/m2
*Not included in analysis because of limited number of postpartum cases.
many such examples in the medical literature, no such index has been applied to obstetrics. Our objective was to develop a risk factor scoring system for the prediction of pregnancy-related thromboembolic events. Methods We conducted a retrospective case-control study of pregnant and postpartum women admitted to the Jack D. Weiler Hospital from 1987 through 1998. Women with a discharge diagnosis indicating thromboembolism formed the study group. For each study subject the 3 women who were delivered immediately before that index patient were selected for the control group. Data collected for each patient included risk factors for the development of thromboembolism. Each risk factor was assigned 1 point, and a total score was then derived for each patient. Risk factors were chosen on the basis of association with thromboembolism, as reported in the medical literature.11, 14, 15 History of previous deep vein thrombosis or pulmonary embolism was assigned a score of 4, because we considered such a history to constitute a more significant risk for a thromboembolic event and wanted to weight it accordingly. The other risk factors that were recorded included the following: age ≥35 years, cesarean delivery, body mass index >29 kg/m2, presence of infection, postpartum oral contraceptive use, non-O blood type, history of previous abdominal or pelvic surgery, fractures, trauma, prolonged bed rest, chronic medical conditions, and personal or familial history of hereditary thrombophilia. A χ2 analysis was used to determine whether a significant association existed between a score of >2 and the
presence of a thromboembolism. P < .05 was considered significant. Results Twenty-one obstetric patients were determined to have had a thromboembolic event (Table I). Most had at least one risk factor, whereas 4 (19%) had no risk factors (or scores of 0). There were 3 case patients with documented thromboembolic events occurring before the index pregnancy. One had had no prophylaxis initiated during the index pregnancy until she was found to have a new clot during the third trimester. Another was found to have a recurrent deep vein thrombosis at the time of diagnosis of the index pregnancy, and she eventually elected to terminate the pregnancy. The final case patient was receiving 5000 units heparin subcutaneously twice each day when a pulmonary embolism developed. Among the 21 cases we found, 19 (90%) occurred antepartum and only 2 (10%) occurred post partum. In the case patients thrombosis occurred during the first trimester in 38%, during the second trimester in 9%, and during the third trimester in 43%. Pulmonary embolisms occurred in 6 case patients (28%). There was only 1 maternal death in this series, that of a morbidly obese woman who had a pulmonary embolism on the first postoperative day after a cesarean delivery. This patient’s risk factors were the mode of delivery and obesity. Each case patient, as well as each control subject, was assigned a score on the basis of the number of risk factors identified during our review of the records. Because of the low number of postpartum thromboembolic events found, statistical analysis of the scoring system could only
Weiss and Bernstein 1075
Volume 182, Number 5 Am J Obstet Gynecol
be applied to the patients with antepartum events. We found that scores >2 were significantly more common among the patients with thromboembolism, with an odds ratio of 4.8 (P = .04; Table II). The sensitivity of a score >2 was 21%, with a specificity of 95% and positive and negative predictive values of 57% and 78%, respectively. Comment Although maternal mortality has decreased dramatically during the last 70 years, pulmonary embolism remains a leading cause of maternal death.14 In light of this fact, a risk factor scoring system such as ours has potential as a useful tool to discriminate patients at high risk for thromboembolism. Our study showed a sensitivity of only 21% for discrimination of patients with thromboembolism during pregnancy. Further refinements to our scoring system, such as adding laboratory screening of women with any risk factor, might improve this system. It is commonly believed that the frequency of thromboembolic events increases as pregnancy advances. We were surprised at our finding that 42% of the antepartum thrombotic events in our series occurred during the first trimester. This finding may simply reflect heightened attention to provision of prophylaxis to high-risk women during the third trimester and post partum. Although previous data suggest that approximately 50% of thromboembolic events occur post partum, we did not find this to be true in our study, in which only 10% of events occurred post partum. This finding may reflect changing practices during the last 50 years in the management of pregnancy. Lengthy hospitalizations and the use of highdose estrogen for lactation suppression—two practices that may have contributed to the incidence of postpartum thromboembolism—have largely been abandoned. Our study had a number of limitations. It was retrospective and featured a small number of case patients. It could not account for symptom-free patients or for highrisk patients who were being treated prophylactically and therefore did not have a thromboembolic episode during the index pregnancy. Our risk factor scoring system could not adequately take into account recent advances in knowledge regarding the etiology of thromboembolism, such as the factor V Leiden mutation. Until recently, testing for most thrombophilic conditions was not widely available, and as far as we know, none of our subjects underwent evaluations for acquired or hereditary thrombophilias at the time of the thromboembolic events. In addition, our method of relying on discharge diagnosis coding as a means of identifying case patients may result in the underascertainment of cases of thromboembolism. We believe that we managed to detect most appropriate case patients, however, because during the study period there were 42,457 deliveries at our institution, and the 21 cases of thromboembolic events that we discovered translate into a rate of
Table II. Risk factor scoring among antepartum patients
Case patients Control subjects TOTAL
Score >2
Score ≤2
Total
4 3 7
15 54 69
19 57 76
Patients with scores >2 were considered at high risk for thromboembolism (P < .05; sensitivity, 21%; specificity, 95%; positive predictive value, 57%; negative predictive value, 78%).
0.5 per 1000 births, a rate that is consistent with other reports.1-3 Finally, we cannot account for other changes in practice during the last decade that may have had an impact on the incidence of thromboembolism in this population. Our hope is that in the future a scoring system such as ours could be used prospectively to determine patients at high risk for thromboembolism. This would facilitate a study featuring random assignment of different methods of prophylaxis. REFERENCES
1. Rutherford SE, Phelan JP. Thromboembolic disease in pregnancy. Clin Perinatol 1986;13:719-39. 2. Aaro LA, Juergens JL. Thrombophlebitis associated with pregnancy. Am J Obstet Gynecol 1971;109:1128-36. 3. Treffers PE, Huidekoper BL, Weenink GH, Kloosterman GJ. Epidemiological observations of thrombo-embolic disease during pregnancy and in the puerperium, in 56,022 women. Int J Gynaecol Obstet 1983;21:327-31. 4. Atrash HK, Koonin LM, Lawson HW, Franks AL, Smith JC. Maternal mortality in the United States 1979-1986. Obstet Gynecol 1990;76:1055-60. 5. Wessler S. Medical management of venous thrombosis. Annu Rev Med 1976;27:313-9. 6. Villasanta U. Thromboembolic disease in pregnancy. Am J Obstet Gynecol 1965;93:142-60. 7. Onyeije CI, Sherer DM, Bernstein PS, Merkatz IR. Peripartum thromboembolism prophylaxis: a survey of practiced methods among perinatologists [abstract]. J Soc Gynecol Investig 1999;6(Suppl):177A. 8. Caprini JA, Arcelus JI, Hasty JH, Tamhane AC, Fabrega F. Clinical assessment of venous thromboembolic risk in surgical patients. Semin Thromb Hemost 1991;17 Suppl 3:304-12. 9. Clayton JK, Anderson JA, McNicol GP. Preoperative prediction of postoperative deep vein thrombosis. BMJ 1976;2:910-2. 10. Crandon AJ, Peel KR, Anderson JA, Thompson V, McNicol GP. Postoperative deep vein thrombosis: identifying high-risk patients. BMJ 1980;7:343-4. 11. Janssen HF, Schachner J, Hubbard J, Hartman JT. The risk of deep vein thrombosis: a computerized epidemiologic approach. Surgery 1987;101:205-12. 12. Sue-Ling HM, Johnston D, McMahon MJ, Philips PR, Davies JA. Pre-operative identification of patients at high risk of deep venous thrombosis after elective major abdominal surgery. Lancet 1986;1:1173-6. 13. Clarke-Pearson DL, DeLong ER, Synan IS, Coleman RE, Creasman WT. Variables associated with postoperative deep venous thrombosis: a prospective study of 411 gynecology patients and creation of a prognostic model. Obstet Gynecol 1987; 69:146-50. 14. Bonnar J. Can more be done in obstetric and gynecologic practice to reduce morbidity and mortality associated with venous thromboembolism? Am J Obstet Gynecol 1999;180:784-91. 15. Jick H, Porter J. Thrombophlebitis of the lower extremities and ABO blood type. Arch Intern Med 1978;138:1566-7.
Key words: Deep vein thrombosis, pregnancy, pulmonary embolism, risk factor scoring, thromboembolism
Thromboembolic events are a major potentially preventable cause of maternal morbidity and mortality, with a quoted incidence of pulmonary embolism during pregnancy between 0.09 and 0.7 per 1000 live births.1-3 Women have a 5-fold increased risk of venous thrombosis during pregnancy,4 and this frequency increases with any additional risk factors.1 Pulmonary embolism will develop in as many as 24% of patients with untreated deep vein thrombosis, resulting in a mortality rate of approximately 13%.5 Among patients with deep vein thromboses who receive proper anticoagulation, pulmonary embolism occurs in 4.5% and the mortality rate is reduced to <1%.6 To date there have been no large randomized trials in the obstetric literature examining the efficacy of thromboembolic prophylaxis during pregnancy and the postpartum period. Estimates of occurrence and recurrence
From the Department of Obstetrics and Gynecology and Women’s Health, Albert Einstein College of Medicine, and Montefiore Medical Center. Reprint requests: Peter S. Bernstein, MD, Jack D. Weiler Hospital of the Albert Einstein College of Medicine, 1825 Eastchester Rd, Room 703, Bronx, NY 10461-2373. Copyright © 2000 by Mosby, Inc. 0002-9378/2000 $12.00 + 0 6/1/105441 doi:10.1067/mob.2000.105441
in pregnancy have not been stratified by risk factors and have been based on retrospective studies. Many of the recommendations from consensus panels regarding prophylactic strategies in pregnancy have been based on trials with nonpregnant patients. A recent survey of maternal-fetal medicine specialists revealed a lack of consensus regarding indications for and use of peripartum thromboembolic prophylaxis.7 This confusion regarding how best to prevent thromboembolism among pregnant and postpartum women reflects the lack of good data to guide management. The development of thromboembolism risk factor scoring systems has been suggested in the surgical and gynecologic literature.8-12 Different indexes have been reported to predict the risk for development of postoperative thromboembolic events in different surgical groups. The application of such indexes allows targeting of appropriate prophylaxis. For example, Caprini et al8 proposed a risk assessment rating system that assigned a predetermined score to each risk factor. On the basis of the overall scores specific modalities for thromboembolic prophylaxis were recommended. Clayton et al9 in the surgical literature and ClarkePearson et al13 in the gynecologic literature have proposed similar prognostic indexes. Although there are 1073
1074 Weiss and Bernstein
May 2000 Am J Obstet Gynecol
Table I. Obstetric patients with thromboembolism and their assigned scores Patient
Thromboembolic event
Gestational age (wk)
Score
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Pulmonary embolism Deep venous thrombosis Deep venous thrombosis Pulmonary embolism Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Deep venous thrombosis Pulmonary embolism Pulmonary embolism Deep venous thrombosis Pulmonary embolism Pulmonary embolism
28 37 28 30 34 12 37 24 12 08 05 11 38 10 09 34 22 30 08 1 d post partum* 5 wk post partum*
4 1 2 0 0 1 1 2 1 1 2 0 0 5 1 3 1 5 1 2 3
Risk factors Previous thrombosis Non-O blood type Age ≥35 y, previous abdominal surgery Body mass index >29.0 kg/m2 Previous abdominal surgery Non-O blood type, body mass index >29.0 kg/m2 Previous abdominal surgery Non-O blood type Non-O blood type, body mass index >29.0 kg/m2 Previous thrombosis, non-O blood type Non-O blood type Non-O blood type, age ≥35 y, sickle cell disease Previous abdominal surgery Previous thrombosis, non-O blood type Non-O blood type Cesarean delivery, body mass index >29.0 kg/m2 Age ≥35 y, non-O blood type, body mass index >29.0 kg/m2
*Not included in analysis because of limited number of postpartum cases.
many such examples in the medical literature, no such index has been applied to obstetrics. Our objective was to develop a risk factor scoring system for the prediction of pregnancy-related thromboembolic events. Methods We conducted a retrospective case-control study of pregnant and postpartum women admitted to the Jack D. Weiler Hospital from 1987 through 1998. Women with a discharge diagnosis indicating thromboembolism formed the study group. For each study subject the 3 women who were delivered immediately before that index patient were selected for the control group. Data collected for each patient included risk factors for the development of thromboembolism. Each risk factor was assigned 1 point, and a total score was then derived for each patient. Risk factors were chosen on the basis of association with thromboembolism, as reported in the medical literature.11, 14, 15 History of previous deep vein thrombosis or pulmonary embolism was assigned a score of 4, because we considered such a history to constitute a more significant risk for a thromboembolic event and wanted to weight it accordingly. The other risk factors that were recorded included the following: age ≥35 years, cesarean delivery, body mass index >29 kg/m2, presence of infection, postpartum oral contraceptive use, non-O blood type, history of previous abdominal or pelvic surgery, fractures, trauma, prolonged bed rest, chronic medical conditions, and personal or familial history of hereditary thrombophilia. A χ2 analysis was used to determine whether a significant association existed between a score of >2 and the
presence of a thromboembolism. P < .05 was considered significant. Results Twenty-one obstetric patients were determined to have had a thromboembolic event (Table I). Most had at least one risk factor, whereas 4 (19%) had no risk factors (or scores of 0). There were 3 case patients with documented thromboembolic events occurring before the index pregnancy. One had had no prophylaxis initiated during the index pregnancy until she was found to have a new clot during the third trimester. Another was found to have a recurrent deep vein thrombosis at the time of diagnosis of the index pregnancy, and she eventually elected to terminate the pregnancy. The final case patient was receiving 5000 units heparin subcutaneously twice each day when a pulmonary embolism developed. Among the 21 cases we found, 19 (90%) occurred antepartum and only 2 (10%) occurred post partum. In the case patients thrombosis occurred during the first trimester in 38%, during the second trimester in 9%, and during the third trimester in 43%. Pulmonary embolisms occurred in 6 case patients (28%). There was only 1 maternal death in this series, that of a morbidly obese woman who had a pulmonary embolism on the first postoperative day after a cesarean delivery. This patient’s risk factors were the mode of delivery and obesity. Each case patient, as well as each control subject, was assigned a score on the basis of the number of risk factors identified during our review of the records. Because of the low number of postpartum thromboembolic events found, statistical analysis of the scoring system could only
Weiss and Bernstein 1075
Volume 182, Number 5 Am J Obstet Gynecol
be applied to the patients with antepartum events. We found that scores >2 were significantly more common among the patients with thromboembolism, with an odds ratio of 4.8 (P = .04; Table II). The sensitivity of a score >2 was 21%, with a specificity of 95% and positive and negative predictive values of 57% and 78%, respectively. Comment Although maternal mortality has decreased dramatically during the last 70 years, pulmonary embolism remains a leading cause of maternal death.14 In light of this fact, a risk factor scoring system such as ours has potential as a useful tool to discriminate patients at high risk for thromboembolism. Our study showed a sensitivity of only 21% for discrimination of patients with thromboembolism during pregnancy. Further refinements to our scoring system, such as adding laboratory screening of women with any risk factor, might improve this system. It is commonly believed that the frequency of thromboembolic events increases as pregnancy advances. We were surprised at our finding that 42% of the antepartum thrombotic events in our series occurred during the first trimester. This finding may simply reflect heightened attention to provision of prophylaxis to high-risk women during the third trimester and post partum. Although previous data suggest that approximately 50% of thromboembolic events occur post partum, we did not find this to be true in our study, in which only 10% of events occurred post partum. This finding may reflect changing practices during the last 50 years in the management of pregnancy. Lengthy hospitalizations and the use of highdose estrogen for lactation suppression—two practices that may have contributed to the incidence of postpartum thromboembolism—have largely been abandoned. Our study had a number of limitations. It was retrospective and featured a small number of case patients. It could not account for symptom-free patients or for highrisk patients who were being treated prophylactically and therefore did not have a thromboembolic episode during the index pregnancy. Our risk factor scoring system could not adequately take into account recent advances in knowledge regarding the etiology of thromboembolism, such as the factor V Leiden mutation. Until recently, testing for most thrombophilic conditions was not widely available, and as far as we know, none of our subjects underwent evaluations for acquired or hereditary thrombophilias at the time of the thromboembolic events. In addition, our method of relying on discharge diagnosis coding as a means of identifying case patients may result in the underascertainment of cases of thromboembolism. We believe that we managed to detect most appropriate case patients, however, because during the study period there were 42,457 deliveries at our institution, and the 21 cases of thromboembolic events that we discovered translate into a rate of
Table II. Risk factor scoring among antepartum patients
Case patients Control subjects TOTAL
Score >2
Score ≤2
Total
4 3 7
15 54 69
19 57 76
Patients with scores >2 were considered at high risk for thromboembolism (P < .05; sensitivity, 21%; specificity, 95%; positive predictive value, 57%; negative predictive value, 78%).
0.5 per 1000 births, a rate that is consistent with other reports.1-3 Finally, we cannot account for other changes in practice during the last decade that may have had an impact on the incidence of thromboembolism in this population. Our hope is that in the future a scoring system such as ours could be used prospectively to determine patients at high risk for thromboembolism. This would facilitate a study featuring random assignment of different methods of prophylaxis. REFERENCES
1. Rutherford SE, Phelan JP. Thromboembolic disease in pregnancy. Clin Perinatol 1986;13:719-39. 2. Aaro LA, Juergens JL. Thrombophlebitis associated with pregnancy. Am J Obstet Gynecol 1971;109:1128-36. 3. Treffers PE, Huidekoper BL, Weenink GH, Kloosterman GJ. Epidemiological observations of thrombo-embolic disease during pregnancy and in the puerperium, in 56,022 women. Int J Gynaecol Obstet 1983;21:327-31. 4. Atrash HK, Koonin LM, Lawson HW, Franks AL, Smith JC. Maternal mortality in the United States 1979-1986. Obstet Gynecol 1990;76:1055-60. 5. Wessler S. Medical management of venous thrombosis. Annu Rev Med 1976;27:313-9. 6. Villasanta U. Thromboembolic disease in pregnancy. Am J Obstet Gynecol 1965;93:142-60. 7. Onyeije CI, Sherer DM, Bernstein PS, Merkatz IR. Peripartum thromboembolism prophylaxis: a survey of practiced methods among perinatologists [abstract]. J Soc Gynecol Investig 1999;6(Suppl):177A. 8. Caprini JA, Arcelus JI, Hasty JH, Tamhane AC, Fabrega F. Clinical assessment of venous thromboembolic risk in surgical patients. Semin Thromb Hemost 1991;17 Suppl 3:304-12. 9. Clayton JK, Anderson JA, McNicol GP. Preoperative prediction of postoperative deep vein thrombosis. BMJ 1976;2:910-2. 10. Crandon AJ, Peel KR, Anderson JA, Thompson V, McNicol GP. Postoperative deep vein thrombosis: identifying high-risk patients. BMJ 1980;7:343-4. 11. Janssen HF, Schachner J, Hubbard J, Hartman JT. The risk of deep vein thrombosis: a computerized epidemiologic approach. Surgery 1987;101:205-12. 12. Sue-Ling HM, Johnston D, McMahon MJ, Philips PR, Davies JA. Pre-operative identification of patients at high risk of deep venous thrombosis after elective major abdominal surgery. Lancet 1986;1:1173-6. 13. Clarke-Pearson DL, DeLong ER, Synan IS, Coleman RE, Creasman WT. Variables associated with postoperative deep venous thrombosis: a prospective study of 411 gynecology patients and creation of a prognostic model. Obstet Gynecol 1987; 69:146-50. 14. Bonnar J. Can more be done in obstetric and gynecologic practice to reduce morbidity and mortality associated with venous thromboembolism? Am J Obstet Gynecol 1999;180:784-91. 15. Jick H, Porter J. Thrombophlebitis of the lower extremities and ABO blood type. Arch Intern Med 1978;138:1566-7.