- Email: [email protected]
Gender bias in use of venous ultrasonography for diagnosis of deep venous thrombosis
Gender bias in use of venous ultrasonography for diagnosis of deep venous thrombosis H u g h G. Beebe, M D , R o b e r t P. Scissons, R V T , Sergio X. Salles-Cunha, P h D , Steven M. Dosick, M D , Ralph C. Whalen, M D , Steven S. Gale, M D , J o h n P. Pigott, M D , and Michael J. Vitti, M D , Toledo, Ohio
Purpose: We observed that ultrasound examinations for deep venous thrombosis (DVT) were more frequently requested for women than for men in our vascular laboratory serving a general outpatient population and referral 774-bed hospital. Because existing literature presents conflicting information about sex differences in occurrence of DVT, we investigated correlation in our population with positive ultrasound study results and risk factors for DVT. Methods: In 13 months, 2055 ultrasound examinations for DVT were requested. O f these, 300 patients (15%) were categorized in four subgroups: 75 ultrasonography-negative men, 75 ultrasonography-negative women, 75 ultrasonography (DVT)-positive men, and 75 ultrasonography (DVT)-positive women for risk factor analysis. Results: Women comprised 64% (I311 of 2055) and men 36% (744 of 2055) of ultrasound examinations requested, but men had significantly higher incidence of DVT-positive ultrasonography results (101 of 744 [14%]) compared with women (118 of 1311 [9%]) (p = 0.002 by chi-square testing). There were no significant sex differences in conventional DVT risk factors and no difference in aggregate number of risk factors. The anatomic distribution of DVT was the same in men as in women. Among those having negative ultrasonography results, significantly more outpatient examinations were performed in women (p = 0.018 by t testing). Conclusions: Gender bias exists in use of ultrasonography for diagnosis of DVT. The greater incidence of women undergoing venous ultrasonography is not explained by higher prevalence of DVT risk factors or of higher occurrence of positive ultrasound examination results. Further investigation is needed to determine whether these differences indicate underuse of ultrasonography in men or overuse in women. (J VAsc SURG 1995;22: 538-42.)
During evaluation o f routine vascular diagnostic laboratory activity, we noted apparent skewed distribution o f venous ultrasound testing in favor o f women. This testing was performed in a hospitalbased laboratory o f a 774-bed tertiary care hospital that also serves a large primary care outpatient activity without any obvious gender bias. On review o f published clinical studies, there appeared to be conflicting evidence o f sex difference in prevalence of deep venous thrombosis (DVT) with approximately equal balance between observations concluding that
From the Jobst Vascular Center, Toledo. Presentedat the SeventhAnnualMeeting of the AmericanVenous Forum, Fort Lauderdale, Fla., Feb. 23-25, 1995. Reprint requests: Hugh G. Beebe, MD, Jobst Vascular Center, 2109 Hughes Dr., Suite 400, Toledo, OH 43606. Copyright © 1995 by The Society for Vascular Surgery and International Societyfor CardiovascularSurgery,North American Chapter. 0741-5214/95/$5.00 + 0 24/6/67248 538
men or women were more commonly afflicted. Thus we sought to understand whether our patient population had unusual characteristics o f particular importance for venous thrombosis that varied by sex. The result of that study forms the basis for this report. METHODS All patients referred to the Vascular Diagnostic Laboratory for lower extremity venous duplex ultrasound examination for D V T from a 13-month interval were included in the study, including both inpatient and outpatient source. Examination was carried out according to standard protocol with current color-flow duplex ultrasound equipment by registered vascular technologists in a laboratory certified by the Intersocietal Commission for Accreditation o f Vascular Laboratories. Primary criterion for positive diagnosis was incompressibility o f a deep vein clearly visualized by
JOURNAL OF VASCULAR SURGERY Volume 22, Number 5
B e e b e e t al.
539
Table I. Distribution of D V T risk factors in study cohort Men
Women
Factor
US +
US -
US +
US -
Sex difference
Mean age (yrs.) Previous D V T Trauma Joint surgery Bed rest Obesity Heart disease Cancer Varices Stroke
61.3 32 1 11 8 34 25 20 23 1
59.5 10 4 11 9 19 14 13 19 0
63.6 24 5 14 10 32 26 24 22 1
59.1 5 2 9 3 21 6 5 24 0
NS NS NS NS NS NS NS NS NS NS
US, Ultrasonography; N S , n o t significant.
B-mode ultrasonography. Secondary criteria included enlarged veins without Doppler spectral evidence of venous flow and presence of increased collateral vessels. Diagnosis was confirmed'by physician review of positive ultrasonography findings in all cases.
After reviewing all patients in the overall study cohort for determination of sex and positive or negative study status, 300 (15%) patients from representative subgroups received detailed evaluation of traditional risk factors for D V T according to the following categories: men or women with negative scanning results and men or women with ultrasound findings diagnostic of DVT. Seventy-five patients from each of these four cohorts were randomly selected. Traditional risk factors for D V T were determined by review of patient records, including patient age, age over 60 years, previous history of documented DVT, recent or past documented pulmonary embolism, trauma or bone fracture within 2 weeks, joint replacement surgery, major abdominal or thoracic cavity surgery within 2 weeks, bed rest or immobilization prolonged beyond 1 week, obesity, congestive heart failure, current diagnosis of malignancy, varicose veins, or history of stroke. Anatomic classification of thrombus location and extent was recorded for the 150 patients with positive scanning results. RESULTS From July 1, 1992 until July 31, 1993, a total of 2055 patients were evaluated for acute D V T of the lower extremity. Women comprised 64% (1311) and men 36% (744) of the total group examined. There was a total of 219 (10.6%) positive examination results resulting in a diagnosis of acute DVT. Women comprised 53.9% (118) and men 46.1% (101) of the positive group. Thus 101 of 744 men (13,6%) but
only 118 of 1311 women (9%) had positive study results, which was a highly significant difference (p = 0.006 by chi squared test). Analysis of 10 traditional risk factors for D V T revealed no statistically significant differences between men and women in the cohorts with positive or negative ultrasound scanning results. The relative importance of these risk factors could be assessed by comparing men and women with a particular risk factor separately against the total subgroup with that risk factor. For example, 62 of 300 (21%) patients of both sexes scanned had a current diagnosis of cancer, of whom 44 (71%) had a positive scanning result (p < 0.001 byttest). However, there was no significant sex distribution difference among 20 men and 24 women with cancer and positive scanning results (p = 0.5). Similarly, among 71 of 300 (24%) patients of both sexes with severe heart disease, incidence of positive scanning results was clearly significant 51 ((72%) (p = 0.03). There was not a difference between men and women within the heart disease subset, however, which was almost equally divided between 25 men and 26 women with positive scanning results (p = 1). The mean age of those of both sexes with positive scanning results was greater than in the groups with negative scanning results, but there was no significant difference between men and women within scanning result cohorts. These results are summarized in Table I. When stratified according to referral source, 300 patients in the detailed analysis cohorts derived from emergency department (n = 36 [12%]), private office outpatient (n = 124 [41%]) and inpatient (n = 140 [47%]). Twenty-seven of 36 patients (75%) of both sexes referred from emergency department source had positive diagnosis of D V T (p = 0.003 by t test), but they were nearly equally divided among 12 men and 15 women. There was
JOURNAL OF VASCULAR SURGERY
540 Beebe et al.
November 1995
Table II. Referral source of study cohort J~elq,
Referral source OP IP ED
US + 26 37 12
WOM~I~
US29 39 7
US + 25 35 15
US44 29 2
Sex difference p = 0.018 NS NS
US, Ultrasonography; OP, outpatient;/P, inpatient; ED, emergency depamnent; NS, not significant.
significant sex difference in referral source among patients having negative duplex examination results. A total of 51 outpatients had positive scanning results nearly equally divided between 26 men and 25 women. O f the remaining 73 outpatients with negative examination results, 29 were men (39%), and 44 were women (59%), which was significant (p = 0.018 by t test) (Table II). Referrals to the vascular diagnostic laboratory originated from a total of a31 physicians, of whom the top 100 requested 89.6% of all tests. The leading specialty designations in terms of volume of tests requested in descending rank order were internal medicine, family practice, emergency physicians, orthopedic surgeons, and vascular surgeons. Women outnumbered men in the hospital inpatient population, 17,953 versus 11,675, during the study interval. This probably represents a bias because of an obstetric service with more than 5000 deliveries annually. Women were more commonly seen in the emergency department, 41,055 women versus 33,945 men, during the same study period. However, the proportionality of women to men seen in the emergency department, 1.2:1, was less skewed than the proportion of women to men having ultrasonography requested for DVT, 1.76:1. DISCUSSION The literature preceding widespread application of imaging studies such as retrograde phlebography or venous duplex ultrasonography seems to have given the impression that women were more subject to DVT than men. Relying largely on history and, to a lesser extent physical examination results, Coon et al) assessed occurrence of DVT in a population of about 8600 patients in the Tecumseh Community Health Study. They found that DVT was significantly skewed toward younger women. Of all 1515 patients of both sexes with DVT, 71 (47%) occurred in women under age 50, but only 8 (5%) occurred among men in this age group. They attributed the higher frequency of DVT and pulmonary embolism in young women to "the added risks.., during
pregnancy and the puerperium." Perhaps this was true in an era when bed rest was believed to be of essential benefit in management of many conditions including recovery from pregnancy. This report associating female sex with venous thromboembolism, which antedated the availability of accurate noninvasive diagnostic imaging, has been cited in other reports and may have contributed to the traditional perspective of many clinicians. The notion that estrogen hormone medication contributes to thrombosis predilection also may be held by many practitioners. A comprehensive review of published reports relevant to this question as of 1992, however, concludes that this is a myth. 2 In an extensive review of the epidemiology and pathogenesis of venous thrombosis, Hirsh et al.3 do not mention sex as a specific direct influence on risk. They do acknowledge that there is a perception of risk associated with pregnancy and the use of birth control pharmaceuticals. However, they conclude "there is no good evidence" of increased DVT risk with low-dose estrogen medication. Franks et alp attempted to study prevalence of venous disease by interrogating 1338 patients living at home by questionnaire. These British National Health Service enrollers were asked questions about venous disease but were not examined objectively for purposes of that study. Female sex had a highly significant association with "phlebitis" (odds ratio, 1.59; 95% confidence interval, 1.16 to 2.22), but only in patients over the age of 50 years. In the United States, with Medicare enrollers computerized records obtained from the Health Care Financing Administration, Kniffin et al.s recently determined that women had a lower incidence rate for pulmonary embolism than men, with adjusted relative risk of 0.86. For DVT the opposite appeared to be true, with women having a significantly increased relative risk of 1.25. This study did not control for the quality of diagnostic evaluation or whether any objective imaging test had been performed. Conversely, Markel et al.6 recently observed that 209 patients with duplex ultrasonography-proven
JOURNAL OF VASCULAR SURGERY Volume 22, Number 5
D V T out of 833 referred for evaluation because of clinical suspicion had a statistically significant (p < 0.05) bias toward male sex. Referral source was not specified in that study. Similarly, Nordstrom et al. 7 evaluated all 366 (36%) patients with phlebographically documented D V T from a total of 1009 who were referred for examination. Referral source was not documented, but this was explicitly a hospital-based population of patients. They found an equally distributed incidence of D V T between men and women of 1.6 per 1000 population per year. Incidence increased markedly above age 50 in men but was not similarly increased in women until a decade later, above age 60. Kierkegaard 8 addressed the interrelationship between age and sex in the incidence of DVT, also in a Swedish population that had a single hospital serving a defined population. The 344 cases with thrombosis were proven by phlebography, and more than 75% of these occurred in outpatients. Beyond age 50, the incidence of D V T was significantly more common in men. Men also constituted a 57% majority of 231 patients with a phlebographlcaliy proven diagnosis from a similarly concentrated population in another Swedish report. 9 Anderson et al. ~0 studied a representative population in the United States. They also found significantly higher rates of both clinically recognized and objectively confirmed D V T in men across all ages up to the ninth decade (p < 0.05). The apparent lack of data sufficient to create a consensus on sex-specific factors is well exemplified by a report analyzing the contributions of 21 factors to D V T risk by constructing a computer program from reports in the literature, n In 20 of 21 factors believed to influence D V T risk, no mention was made of influence by sex except for pregnancy, where there was necessarily a 100% sex-specific increase in risk. In other cardiovascular disease categories sex differences in diagnostic and therapeutic procedures have been examined, most notably coronary artery disease. Ultrasound diagnostic procedures such as echocardiography are more likely to be accomplished in women than in men. 12 However, the burden of evidence strongly suggests that women have been excluded or ignored in major trials of coronary artery disease evaluation, prevention, and treatment, a3 Analysis of hospital treatment of non-coronary artery disease has, however, shown a lack of sexual bias in clinical decision making during hospitalization. 14 It appears uncertain whether women are more expressive of complaints that might suggest the protean or subtle qualities of D V T that lead clinicians
Beebe et al.
541
to accurately assess the possibility of D V T by ultrasound imaging. Verbrugge and Steiner ts from the University of Michigan examined patterns from the 1975 Ambulatory Medical Care Survey and concluded that women receive generally more medical care than men do even though severity of illness differences were not shown in that study. Others have determined after detailed study that little or no data exist from actual patient/doctor encounters whether quantity and types of symptoms reported are different between men and women. 16 Katz et al.~7 looked for venous hemodynamics changes as a consequence of daily activity in normal volunteers. Although 20% of their small study cohort appeared to have deterioration suggestive of leg valve dysfunction, this did not have gender differences. 17 Wille-Jorgensen et al.18 made a stimulating and provocative observation by comparing subjective and objective evaluation of postphlebitic syndrome in 3608 patients in 1991. Although subjective perception of this venous disorder was much more frequent than objective evidence, they found this was significantly more common in women than in men at all ages. Among 1843 women, subjective perception of postphlebitic venous disease was twice as common as objective evidence. Although this margin declined with age, it was significantly greater through the sixth decade. Cogo et al.19 reported a very interesting study bearing on perceptions of D V T risk by evaluating 426 consecutive outpatients with clinically suspected D V T and no prior venous thromboembolic history. Physician specialists prospectively evaluated risk factors before venography in all patients. In a group of 278 judged to have "idiopathy" because they lacked risk factors of trauma, major surgery, or prolonged immobilization, 90 (32%) proved to have DVT. There was a highly significant sex difference skewing the population referred for evaluation. One hundred twenty-two of the 188 (65%) who did not have D V T were women, but only 37 of the 90 proven D V T cases (41%) were women (p < 0.00i by chi-squared testing). A recent review of the epidemiology of venous thrombosis by Carter 2° concludes that further epidemiologic study is needed especially on incidence in the general community. Cost-effectiveness of diagnostic studies cannot be accurately evaluated if unexplained bias in referral of any sort skews results. In our study population, which may be representative of the type of hospital-based vascular laboratory that is becoming the norm, traditional risk factors were equally distributed among men and women. Yet there was a significant gender bias in favor of more women being referred for examination. We conclude
542
JOURNAL OF VASCULARSURGERY November 1995
Beebe et al.
t h a t g e n d e r bias i n D V T e v a l u a t i o n , as d e m o n s t r a t e d i n this study, is e v i d e n c e o f t h e n e e d for f u r t h e r study.
REFERENCES 1. Coon WW, Willis PW III, Keller lB. Venous thromboembolism and other venous disease in the Tecumseh Community Health Study. Circulation 1973;48:839-46. 2. Wren BG. The effect ofoestrogen on the female cardiovascular system. Med I Aust 1992;157:204-8. 3. Hirsh J, Hull RD, Raskob GE. Epidemiology and pathogenesis of venous thrombosis. JAm Colt Cardiol 1986;8:104B13B. 4. Franks PJ, Wright DDI, Moffatt CJ, et al. Prevalence of venous disease: a community study in West London. Eur J Surg 1992;158:143-7. 5. Kniffin WD Jr, Baron JA, Barrett J, et al. The epidemiology of diagnosed pulmonary embolism and deep venous thrombosis in the elderly. Arch Intern Med 1994;I54:861-6. 6. Markel A, Manzo RA, Bergelin RO, et al. Pattern and distribution of thrombi in acute venous thrombosis. Arch Surg 1992;127:305-9. 7. Nordstrom M, Lindblad B, Bergqvist D, et al. A prospective study of the incidence of deep-vein thrombosis within a defined urban population. J Int Med 1992;232:155-60. 8. Kierkegaard A. Incidence of acute deep vein thrombosis in two districts. Acta Chir Scand 1980;146:267-9. 9. Nylander G, Olivecrona H. The phlebographic pattern of acute leg thrombosis within a defined urban population. Acta Chir Scand 1976;142:505-Ii. i0. Anderson F A l r , Wheeler HB, Goldberg RJ, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary function: the Worcester DVT Study. Arch Intern Med 199I;15I:933-8.
11. Janssen HE, Schachner J, Hubbard J, et al. The risk of deep venous thrombosis: a computerized epidemiologic approach. Surgery 1987;I01:205-12. 12. Chiriboga DE, Yarzebski J, Goldberg RJ, et al. A communitywide perspective of gender differences and temporal trends in the use of diagnostic and revascularization procedures for acute myocardial infarction. Am I Cardiol 1993;71:268-73. 13. Kuhn FE, Rackley CE. Coronary artery disease in women: risk factors, evaluation, treatment, and prevention. Arch Intern Med 1993;153:2626-36. 14. Pearson ML, Kahn KL, Harrison ER, et al. Differences in quality of care for hospitalized elderly men and women. JAMA 1992;268: I883-9. 15. Verbrugge LM, Steiner RP. Physician treatment of men and women patients. Med Care 1981;19:609-32. 16. Weisman CS, Teitelbaum MA. Women and health care communication. Patient Educ Couns 1989;13:183-99. 17. Katz ML, Comerota AJ, Kerr RP, et al. Variability of venous hemodynamics with daily activity. I VASe SURG 1994;19: 361-5. 18. Wille-Jorgensen P, }'orgensen T, Andersen M, et al. Postphlebitic syndrome and general surgery: an epidemiologic investigation. Angiology 1991;42:397-403. 19. Cogo A, Bernardi E, Prandoni P, et al. Acquired risk factors for deep-vein thrombosis in symptomatic outpatients. Arch Intern Med 1994; 154: I64-8. 20. Carter CJ. The natural history and epidemiology of venous thrombosis. Prog Cardiovas Dis 1994;3:423-38.
Submitted March 17, i995; accepted lune 14, I995.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNAL OF VASCULAR SURGERY for 1995 are available to subscribers only. They may be purchased from the publisher at a cost of $84.00 for domestic, $109.14 for Canadian, and $102.00 for international subscribers for Vol. 21 (January to June) and Vol. 22 (July to December). Price includes shipping charges. Each b o u n d volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Subscription Services, Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, M O 63146-3318, USA. In the U n i t e d States call toll free (800)325-4177, ext. 4351. In Missouri or foreign countries call (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.
Purpose: We observed that ultrasound examinations for deep venous thrombosis (DVT) were more frequently requested for women than for men in our vascular laboratory serving a general outpatient population and referral 774-bed hospital. Because existing literature presents conflicting information about sex differences in occurrence of DVT, we investigated correlation in our population with positive ultrasound study results and risk factors for DVT. Methods: In 13 months, 2055 ultrasound examinations for DVT were requested. O f these, 300 patients (15%) were categorized in four subgroups: 75 ultrasonography-negative men, 75 ultrasonography-negative women, 75 ultrasonography (DVT)-positive men, and 75 ultrasonography (DVT)-positive women for risk factor analysis. Results: Women comprised 64% (I311 of 2055) and men 36% (744 of 2055) of ultrasound examinations requested, but men had significantly higher incidence of DVT-positive ultrasonography results (101 of 744 [14%]) compared with women (118 of 1311 [9%]) (p = 0.002 by chi-square testing). There were no significant sex differences in conventional DVT risk factors and no difference in aggregate number of risk factors. The anatomic distribution of DVT was the same in men as in women. Among those having negative ultrasonography results, significantly more outpatient examinations were performed in women (p = 0.018 by t testing). Conclusions: Gender bias exists in use of ultrasonography for diagnosis of DVT. The greater incidence of women undergoing venous ultrasonography is not explained by higher prevalence of DVT risk factors or of higher occurrence of positive ultrasound examination results. Further investigation is needed to determine whether these differences indicate underuse of ultrasonography in men or overuse in women. (J VAsc SURG 1995;22: 538-42.)
During evaluation o f routine vascular diagnostic laboratory activity, we noted apparent skewed distribution o f venous ultrasound testing in favor o f women. This testing was performed in a hospitalbased laboratory o f a 774-bed tertiary care hospital that also serves a large primary care outpatient activity without any obvious gender bias. On review o f published clinical studies, there appeared to be conflicting evidence o f sex difference in prevalence of deep venous thrombosis (DVT) with approximately equal balance between observations concluding that
From the Jobst Vascular Center, Toledo. Presentedat the SeventhAnnualMeeting of the AmericanVenous Forum, Fort Lauderdale, Fla., Feb. 23-25, 1995. Reprint requests: Hugh G. Beebe, MD, Jobst Vascular Center, 2109 Hughes Dr., Suite 400, Toledo, OH 43606. Copyright © 1995 by The Society for Vascular Surgery and International Societyfor CardiovascularSurgery,North American Chapter. 0741-5214/95/$5.00 + 0 24/6/67248 538
men or women were more commonly afflicted. Thus we sought to understand whether our patient population had unusual characteristics o f particular importance for venous thrombosis that varied by sex. The result of that study forms the basis for this report. METHODS All patients referred to the Vascular Diagnostic Laboratory for lower extremity venous duplex ultrasound examination for D V T from a 13-month interval were included in the study, including both inpatient and outpatient source. Examination was carried out according to standard protocol with current color-flow duplex ultrasound equipment by registered vascular technologists in a laboratory certified by the Intersocietal Commission for Accreditation o f Vascular Laboratories. Primary criterion for positive diagnosis was incompressibility o f a deep vein clearly visualized by
JOURNAL OF VASCULAR SURGERY Volume 22, Number 5
B e e b e e t al.
539
Table I. Distribution of D V T risk factors in study cohort Men
Women
Factor
US +
US -
US +
US -
Sex difference
Mean age (yrs.) Previous D V T Trauma Joint surgery Bed rest Obesity Heart disease Cancer Varices Stroke
61.3 32 1 11 8 34 25 20 23 1
59.5 10 4 11 9 19 14 13 19 0
63.6 24 5 14 10 32 26 24 22 1
59.1 5 2 9 3 21 6 5 24 0
NS NS NS NS NS NS NS NS NS NS
US, Ultrasonography; N S , n o t significant.
B-mode ultrasonography. Secondary criteria included enlarged veins without Doppler spectral evidence of venous flow and presence of increased collateral vessels. Diagnosis was confirmed'by physician review of positive ultrasonography findings in all cases.
After reviewing all patients in the overall study cohort for determination of sex and positive or negative study status, 300 (15%) patients from representative subgroups received detailed evaluation of traditional risk factors for D V T according to the following categories: men or women with negative scanning results and men or women with ultrasound findings diagnostic of DVT. Seventy-five patients from each of these four cohorts were randomly selected. Traditional risk factors for D V T were determined by review of patient records, including patient age, age over 60 years, previous history of documented DVT, recent or past documented pulmonary embolism, trauma or bone fracture within 2 weeks, joint replacement surgery, major abdominal or thoracic cavity surgery within 2 weeks, bed rest or immobilization prolonged beyond 1 week, obesity, congestive heart failure, current diagnosis of malignancy, varicose veins, or history of stroke. Anatomic classification of thrombus location and extent was recorded for the 150 patients with positive scanning results. RESULTS From July 1, 1992 until July 31, 1993, a total of 2055 patients were evaluated for acute D V T of the lower extremity. Women comprised 64% (1311) and men 36% (744) of the total group examined. There was a total of 219 (10.6%) positive examination results resulting in a diagnosis of acute DVT. Women comprised 53.9% (118) and men 46.1% (101) of the positive group. Thus 101 of 744 men (13,6%) but
only 118 of 1311 women (9%) had positive study results, which was a highly significant difference (p = 0.006 by chi squared test). Analysis of 10 traditional risk factors for D V T revealed no statistically significant differences between men and women in the cohorts with positive or negative ultrasound scanning results. The relative importance of these risk factors could be assessed by comparing men and women with a particular risk factor separately against the total subgroup with that risk factor. For example, 62 of 300 (21%) patients of both sexes scanned had a current diagnosis of cancer, of whom 44 (71%) had a positive scanning result (p < 0.001 byttest). However, there was no significant sex distribution difference among 20 men and 24 women with cancer and positive scanning results (p = 0.5). Similarly, among 71 of 300 (24%) patients of both sexes with severe heart disease, incidence of positive scanning results was clearly significant 51 ((72%) (p = 0.03). There was not a difference between men and women within the heart disease subset, however, which was almost equally divided between 25 men and 26 women with positive scanning results (p = 1). The mean age of those of both sexes with positive scanning results was greater than in the groups with negative scanning results, but there was no significant difference between men and women within scanning result cohorts. These results are summarized in Table I. When stratified according to referral source, 300 patients in the detailed analysis cohorts derived from emergency department (n = 36 [12%]), private office outpatient (n = 124 [41%]) and inpatient (n = 140 [47%]). Twenty-seven of 36 patients (75%) of both sexes referred from emergency department source had positive diagnosis of D V T (p = 0.003 by t test), but they were nearly equally divided among 12 men and 15 women. There was
JOURNAL OF VASCULAR SURGERY
540 Beebe et al.
November 1995
Table II. Referral source of study cohort J~elq,
Referral source OP IP ED
US + 26 37 12
WOM~I~
US29 39 7
US + 25 35 15
US44 29 2
Sex difference p = 0.018 NS NS
US, Ultrasonography; OP, outpatient;/P, inpatient; ED, emergency depamnent; NS, not significant.
significant sex difference in referral source among patients having negative duplex examination results. A total of 51 outpatients had positive scanning results nearly equally divided between 26 men and 25 women. O f the remaining 73 outpatients with negative examination results, 29 were men (39%), and 44 were women (59%), which was significant (p = 0.018 by t test) (Table II). Referrals to the vascular diagnostic laboratory originated from a total of a31 physicians, of whom the top 100 requested 89.6% of all tests. The leading specialty designations in terms of volume of tests requested in descending rank order were internal medicine, family practice, emergency physicians, orthopedic surgeons, and vascular surgeons. Women outnumbered men in the hospital inpatient population, 17,953 versus 11,675, during the study interval. This probably represents a bias because of an obstetric service with more than 5000 deliveries annually. Women were more commonly seen in the emergency department, 41,055 women versus 33,945 men, during the same study period. However, the proportionality of women to men seen in the emergency department, 1.2:1, was less skewed than the proportion of women to men having ultrasonography requested for DVT, 1.76:1. DISCUSSION The literature preceding widespread application of imaging studies such as retrograde phlebography or venous duplex ultrasonography seems to have given the impression that women were more subject to DVT than men. Relying largely on history and, to a lesser extent physical examination results, Coon et al) assessed occurrence of DVT in a population of about 8600 patients in the Tecumseh Community Health Study. They found that DVT was significantly skewed toward younger women. Of all 1515 patients of both sexes with DVT, 71 (47%) occurred in women under age 50, but only 8 (5%) occurred among men in this age group. They attributed the higher frequency of DVT and pulmonary embolism in young women to "the added risks.., during
pregnancy and the puerperium." Perhaps this was true in an era when bed rest was believed to be of essential benefit in management of many conditions including recovery from pregnancy. This report associating female sex with venous thromboembolism, which antedated the availability of accurate noninvasive diagnostic imaging, has been cited in other reports and may have contributed to the traditional perspective of many clinicians. The notion that estrogen hormone medication contributes to thrombosis predilection also may be held by many practitioners. A comprehensive review of published reports relevant to this question as of 1992, however, concludes that this is a myth. 2 In an extensive review of the epidemiology and pathogenesis of venous thrombosis, Hirsh et al.3 do not mention sex as a specific direct influence on risk. They do acknowledge that there is a perception of risk associated with pregnancy and the use of birth control pharmaceuticals. However, they conclude "there is no good evidence" of increased DVT risk with low-dose estrogen medication. Franks et alp attempted to study prevalence of venous disease by interrogating 1338 patients living at home by questionnaire. These British National Health Service enrollers were asked questions about venous disease but were not examined objectively for purposes of that study. Female sex had a highly significant association with "phlebitis" (odds ratio, 1.59; 95% confidence interval, 1.16 to 2.22), but only in patients over the age of 50 years. In the United States, with Medicare enrollers computerized records obtained from the Health Care Financing Administration, Kniffin et al.s recently determined that women had a lower incidence rate for pulmonary embolism than men, with adjusted relative risk of 0.86. For DVT the opposite appeared to be true, with women having a significantly increased relative risk of 1.25. This study did not control for the quality of diagnostic evaluation or whether any objective imaging test had been performed. Conversely, Markel et al.6 recently observed that 209 patients with duplex ultrasonography-proven
JOURNAL OF VASCULAR SURGERY Volume 22, Number 5
D V T out of 833 referred for evaluation because of clinical suspicion had a statistically significant (p < 0.05) bias toward male sex. Referral source was not specified in that study. Similarly, Nordstrom et al. 7 evaluated all 366 (36%) patients with phlebographically documented D V T from a total of 1009 who were referred for examination. Referral source was not documented, but this was explicitly a hospital-based population of patients. They found an equally distributed incidence of D V T between men and women of 1.6 per 1000 population per year. Incidence increased markedly above age 50 in men but was not similarly increased in women until a decade later, above age 60. Kierkegaard 8 addressed the interrelationship between age and sex in the incidence of DVT, also in a Swedish population that had a single hospital serving a defined population. The 344 cases with thrombosis were proven by phlebography, and more than 75% of these occurred in outpatients. Beyond age 50, the incidence of D V T was significantly more common in men. Men also constituted a 57% majority of 231 patients with a phlebographlcaliy proven diagnosis from a similarly concentrated population in another Swedish report. 9 Anderson et al. ~0 studied a representative population in the United States. They also found significantly higher rates of both clinically recognized and objectively confirmed D V T in men across all ages up to the ninth decade (p < 0.05). The apparent lack of data sufficient to create a consensus on sex-specific factors is well exemplified by a report analyzing the contributions of 21 factors to D V T risk by constructing a computer program from reports in the literature, n In 20 of 21 factors believed to influence D V T risk, no mention was made of influence by sex except for pregnancy, where there was necessarily a 100% sex-specific increase in risk. In other cardiovascular disease categories sex differences in diagnostic and therapeutic procedures have been examined, most notably coronary artery disease. Ultrasound diagnostic procedures such as echocardiography are more likely to be accomplished in women than in men. 12 However, the burden of evidence strongly suggests that women have been excluded or ignored in major trials of coronary artery disease evaluation, prevention, and treatment, a3 Analysis of hospital treatment of non-coronary artery disease has, however, shown a lack of sexual bias in clinical decision making during hospitalization. 14 It appears uncertain whether women are more expressive of complaints that might suggest the protean or subtle qualities of D V T that lead clinicians
Beebe et al.
541
to accurately assess the possibility of D V T by ultrasound imaging. Verbrugge and Steiner ts from the University of Michigan examined patterns from the 1975 Ambulatory Medical Care Survey and concluded that women receive generally more medical care than men do even though severity of illness differences were not shown in that study. Others have determined after detailed study that little or no data exist from actual patient/doctor encounters whether quantity and types of symptoms reported are different between men and women. 16 Katz et al.~7 looked for venous hemodynamics changes as a consequence of daily activity in normal volunteers. Although 20% of their small study cohort appeared to have deterioration suggestive of leg valve dysfunction, this did not have gender differences. 17 Wille-Jorgensen et al.18 made a stimulating and provocative observation by comparing subjective and objective evaluation of postphlebitic syndrome in 3608 patients in 1991. Although subjective perception of this venous disorder was much more frequent than objective evidence, they found this was significantly more common in women than in men at all ages. Among 1843 women, subjective perception of postphlebitic venous disease was twice as common as objective evidence. Although this margin declined with age, it was significantly greater through the sixth decade. Cogo et al.19 reported a very interesting study bearing on perceptions of D V T risk by evaluating 426 consecutive outpatients with clinically suspected D V T and no prior venous thromboembolic history. Physician specialists prospectively evaluated risk factors before venography in all patients. In a group of 278 judged to have "idiopathy" because they lacked risk factors of trauma, major surgery, or prolonged immobilization, 90 (32%) proved to have DVT. There was a highly significant sex difference skewing the population referred for evaluation. One hundred twenty-two of the 188 (65%) who did not have D V T were women, but only 37 of the 90 proven D V T cases (41%) were women (p < 0.00i by chi-squared testing). A recent review of the epidemiology of venous thrombosis by Carter 2° concludes that further epidemiologic study is needed especially on incidence in the general community. Cost-effectiveness of diagnostic studies cannot be accurately evaluated if unexplained bias in referral of any sort skews results. In our study population, which may be representative of the type of hospital-based vascular laboratory that is becoming the norm, traditional risk factors were equally distributed among men and women. Yet there was a significant gender bias in favor of more women being referred for examination. We conclude
542
JOURNAL OF VASCULARSURGERY November 1995
Beebe et al.
t h a t g e n d e r bias i n D V T e v a l u a t i o n , as d e m o n s t r a t e d i n this study, is e v i d e n c e o f t h e n e e d for f u r t h e r study.
REFERENCES 1. Coon WW, Willis PW III, Keller lB. Venous thromboembolism and other venous disease in the Tecumseh Community Health Study. Circulation 1973;48:839-46. 2. Wren BG. The effect ofoestrogen on the female cardiovascular system. Med I Aust 1992;157:204-8. 3. Hirsh J, Hull RD, Raskob GE. Epidemiology and pathogenesis of venous thrombosis. JAm Colt Cardiol 1986;8:104B13B. 4. Franks PJ, Wright DDI, Moffatt CJ, et al. Prevalence of venous disease: a community study in West London. Eur J Surg 1992;158:143-7. 5. Kniffin WD Jr, Baron JA, Barrett J, et al. The epidemiology of diagnosed pulmonary embolism and deep venous thrombosis in the elderly. Arch Intern Med 1994;I54:861-6. 6. Markel A, Manzo RA, Bergelin RO, et al. Pattern and distribution of thrombi in acute venous thrombosis. Arch Surg 1992;127:305-9. 7. Nordstrom M, Lindblad B, Bergqvist D, et al. A prospective study of the incidence of deep-vein thrombosis within a defined urban population. J Int Med 1992;232:155-60. 8. Kierkegaard A. Incidence of acute deep vein thrombosis in two districts. Acta Chir Scand 1980;146:267-9. 9. Nylander G, Olivecrona H. The phlebographic pattern of acute leg thrombosis within a defined urban population. Acta Chir Scand 1976;142:505-Ii. i0. Anderson F A l r , Wheeler HB, Goldberg RJ, et al. A population-based perspective of the hospital incidence and case-fatality rates of deep vein thrombosis and pulmonary function: the Worcester DVT Study. Arch Intern Med 199I;15I:933-8.
11. Janssen HE, Schachner J, Hubbard J, et al. The risk of deep venous thrombosis: a computerized epidemiologic approach. Surgery 1987;I01:205-12. 12. Chiriboga DE, Yarzebski J, Goldberg RJ, et al. A communitywide perspective of gender differences and temporal trends in the use of diagnostic and revascularization procedures for acute myocardial infarction. Am I Cardiol 1993;71:268-73. 13. Kuhn FE, Rackley CE. Coronary artery disease in women: risk factors, evaluation, treatment, and prevention. Arch Intern Med 1993;153:2626-36. 14. Pearson ML, Kahn KL, Harrison ER, et al. Differences in quality of care for hospitalized elderly men and women. JAMA 1992;268: I883-9. 15. Verbrugge LM, Steiner RP. Physician treatment of men and women patients. Med Care 1981;19:609-32. 16. Weisman CS, Teitelbaum MA. Women and health care communication. Patient Educ Couns 1989;13:183-99. 17. Katz ML, Comerota AJ, Kerr RP, et al. Variability of venous hemodynamics with daily activity. I VASe SURG 1994;19: 361-5. 18. Wille-Jorgensen P, }'orgensen T, Andersen M, et al. Postphlebitic syndrome and general surgery: an epidemiologic investigation. Angiology 1991;42:397-403. 19. Cogo A, Bernardi E, Prandoni P, et al. Acquired risk factors for deep-vein thrombosis in symptomatic outpatients. Arch Intern Med 1994; 154: I64-8. 20. Carter CJ. The natural history and epidemiology of venous thrombosis. Prog Cardiovas Dis 1994;3:423-38.
Submitted March 17, i995; accepted lune 14, I995.
BOUND VOLUMES AVAILABLE TO SUBSCRIBERS Bound volumes of the JOURNAL OF VASCULAR SURGERY for 1995 are available to subscribers only. They may be purchased from the publisher at a cost of $84.00 for domestic, $109.14 for Canadian, and $102.00 for international subscribers for Vol. 21 (January to June) and Vol. 22 (July to December). Price includes shipping charges. Each b o u n d volume contains a subject and author index, and all advertising is removed. Copies are shipped within 60 days after publication of the last issue in the volume. The binding is durable buckram with the journal name, volume number, and year stamped in gold on the spine. Payment must accompany all orders. Contact Subscription Services, Mosby-Year Book, Inc., 11830 Westline Industrial Dr., St. Louis, M O 63146-3318, USA. In the U n i t e d States call toll free (800)325-4177, ext. 4351. In Missouri or foreign countries call (314)453-4351. Subscriptions must be in force to qualify. Bound volumes are not available in place of a regular JOURNAL subscription.