Clinically Important Pulmonary Emboli: Does Calf Vein US Alter Outcomes? 1

Ultrasonography Ronald H. Gottlieb, MD Jani Widjaja, MD Sonu Mehra, MD William B. Robinette, MS, RDMS, RVT Clinically Important Pulmonary Emboli: Does Calf Vein US Alter Outcomes? 1 Index terms: Embolism, pulmonary, 60.72 Extremities, thrombosis, 93.77 Extremities, US, 93.1298 Radiology 1999; 211:25 29 Abbreviations: DVT deep venous thrombosis PE pulmonary embolism 1 From the Department of Radiology, University of Rochester Medical Center, 601 Elmwood Ave, Box 648, Rochester, NY 14642. Received June 18, 1998; revision requested July 28; revision received August 19; accepted October 7. Address reprint requests to R.H.G. RSNA, 1999 PURPOSE: To assess whether calf imaging is necessary to identify patients at risk for developing clinically important pulmonary emboli (PE) or propagation of calf deep venous thrombosis (DVT) when the initial thigh ultrasonographic (US) scan is negative for DVT. MATERIALS AND METHODS: The authors retrospectively evaluated the radiology reports from 283 patients (168 female and 115 male patients; mean age, 55.7 years; age range, 1 93 years) in whom US was performed to rule out lower extremity DVT. In all patients, the initial thigh examination was negative for DVT. All patients were classified as to the reason for the examination, risk factors for DVT (including recent surgery), whether they received anticoagulation therapy, and findings on calf US scans. Adverse outcomes were considered a clinically important PE or DVT in the thigh. RESULTS: Only 1.1% of patients (95% CI 0.2%, 3.1%) had adverse outcomes. Adverse outcomes occurred only in postsurgical patients (P.028) and were not related to the presence or absence of calf DVT or method of treatment. CONCLUSION: US of the calf is unnecessary at initial evaluation to identify patients at risk of clinically important PE or propagation of DVT into the thigh. Author contributions: Guarantor of integrity of entire study, R.H.G.; study concepts and design, R.H.G.; definition of intellectual content, R.H.G.; literature research, R.H.G., W.B.R.; data acquisition, J.W., S.M.; data analysis, R.H.G., W.B.R., S.M.; statistical analysis, R.H.G.; manuscript preparation and editing, R.H.G.; manuscript review, R.H.G., J.W. Pulmonary embolism (PE) was estimated to be the cause of death in approximately 30,000 patients in 1982 and was diagnosed 120,000 times in 1985 (1). The clinical challenge has been to identify patients at risk for PE to reduce the morbidity and mortality of this disease process. Efforts have been directed toward diagnosing and treating thrombi in the deep venous system of the lower extremities because most emboli to the lungs arise from these veins (2). Ultrasonography (US) has largely replaced conventional venography as the initial diagnostic test for suspected lower extremity venous thrombosis because it is noninvasive and iodinated contrast media or ionizing radiation is not used. US, however, has been found to be less accurate than conventional venography for diagnosing deep venous thrombosis (DVT) in the calf veins (3 6). The importance of isolated calf vein DVT as the cause of clinically important PE or persistent lower extremity symptoms has been the subject of extensive debate in the literature (7 16). Adoption of a policy of scanning only the calf in patients with localized symptoms or physical findings would result in substantial time savings for many US laboratories. In our laboratory, where we routinely scan the calf in patients suspected of having DVT, a substantial number of calf vein evaluations are indeterminate because, frequently, all of the deep veins of the calf are not visualized. This can lead, depending on clinical suspicion, to either follow-up US or conventional venography. Conventional venography and follow-up US would be unnecessary if there were no substantial risk of patients with negative thigh Doppler US scans developing PE or femoropopliteal thrombosis. The elimination of conventional venography in this setting would reduce patient morbidity (reactions to contrast material, thrombophlebitis). The reduction in subsequent testing (follow-up US and conventional venography) would lower the cost of care. Alternatively, conventional venography and follow-up US would be necessary if a substantial number of patients with negative thigh US scans develop PE or femoropopliteal thrombosis (which places the patient at increased risk of PE). 25

The purpose of our study was to evaluate whether US of the calf is necessary to detect patients at increased risk for the subsequent development of adverse outcomes, which we considered to be clinically important PE or propagation of DVT into the thigh. Because calf US was performed in all our patients, we could correlate these results with patient outcomes and treatment decisions. MATERIALS AND METHODS We retrospectively reviewed the radiology reports of all lower extremity US scans obtained in patients referred to our laboratory to exclude a DVT between August 1, 1996, and March 30, 1997 (n 397). All patients underwent US of both the thigh and calf. We excluded patients with DVT identified in the thigh and patients with indeterminate studies of the thigh. We did not exclude patients with a history of DVT or isolated thrombosis of a superficial vein in the thigh or calf. We followed up all patients in whom the initial US scan was negative for DVT (n 284) to determine whether they subsequently developed PE or thigh DVT. Follow-up was performed by means of hospital medical record review, a telephone call to either the patient or his or her physician, and review of the Rochester Hospital Inpatient Database. The Rochester Hospital Inpatient Database lists the International Classification of Diseases (17) clinical modification codes for the discharge diagnoses at the six major hospitals in the county (population of approximately 750,000 people) where our university hospital is located. We also reviewed, by using our radiology information system, all imaging studies (including subsequent leg US scans, conventional venograms, nuclear medicine lung scans, and pulmonary arteriograms) obtained in the patients at our institution. Follow-up was for at least 6 months for the 256 patients who survived 6 months or longer after their initial US scan. Twenty-eight patients died less than 6 months after their initial US study from a cause other than PE. The cause of death was determined by reviewing the death certificates or medical records in 27 patients and an autopsy report in one patient. Local hospital discharge diagnoses were available for all patients. A telephone interview with the patient or his or her physician, a review of the medical records, or both were considered to be adequate follow-up (Table 1). We received Institutional Review Board approval to conduct telephone interviews with the patients. Informed consent was not considered to be necessary by our Institutional Review Board because no additional testing was done as a result of our study. One patient was excluded because we were unable to obtain adequate follow-up with either a review of the medical records or a telephone interview. Thus, our study consisted of 283 patients (168 female and 115 male patients; mean age, 55.7 years; age range, 1 93 years) who underwent 392 baseline limb US evaluations (174 unilateral examinations, 109 bilateral examinations). A baseline examination was considered to be the initial lower extremity US examination the patient underwent during the study period; all subsequent lower extremity US examinations were considered follow-up studies. The beginning of the follow-up period was the time of the baseline US study. Adverse outcomes were considered to be the development of clinically important PE or thigh DVT. Clinically important PE were those in which the patient presented with symptoms (eg, shortness of breath, chest pain) that motivated testing to exclude a PE or contributed to the patient s death. Silent PE could not be detected with our method of follow-up because all patients were not tested to exclude a PE after the baseline US study. A PE was considered to be present with either a high-probability nuclear medicine lung scan or a positive pulmonary arteriogram. A thigh DVT had to be documented with either conventional venography or a subsequent Doppler US study. Patients were classified according to the indication for the US examination (Table 1), as follows: (a) symptoms or physical findings related to at least one of the lower extremities; (b) PE suspected on the basis of symptoms, physical findings, or laboratory test results; or (c) surveillance in high-risk patients for the development of lower extremity DVT or PE (eg, postsurgical patients). We also recorded pertinent risk factors for the development of either DVT or PE (Table 1), including current malignancy, recent surgery (within 2 months of baseline US), previous leg DVT, trauma requiring hospitalization, and a history of birth control pills. Patients were considered adequately treated for DVT or PE if they received intravenous heparin followed by a course of oral warfarin or not adequately treated if no anticoagulation therapy or less aggressive anticoagulation therapy was administered (eg, aspirin or subcutaneous heparin only). TABLE 1 Patient Parameters No. of Patients Parameter (n 283) Method of follow-up Telephone call* 3 (1.1) Medical record 42 (14.8) Both 238 (84.1) Indications for US Leg symptoms or physical findings 174 (61.5) Suspected PE 47 (16.6) Suspected PE and leg symptoms or physical findings 22 (7.8) Surveillance in high-risk patients 21 (7.4) Risk factors for PE or DVT Malignancy 57 (20.1) Recent surgery 87 (30.7) Trauma 31 (11.0) Previous DVT 28 (9.9) Use of birth control pills 4 (1.4) Note. Numbers in parentheses are percentages. Local hospital discharge diagnoses were reviewed for all patients to see if any patient was discharged with a diagnosis of lowerextremity DVT or PE. Follow-up was for at least 6 months. * The patient, or a physician who was involved in treating the patient, was called specifically to inquire about whether the patient experienced lower extremity symptoms or developed PE or lower extremity DVT following the initial US study. These patients were generally those at increased risk for PE or DVT because they were immobile (eg, postsurgical patients). TABLE 2 Correlation of Calf US Results with Treatment and Patient Outcome Findings at Treatment* Calf US Yes No Negative 4 (1) 22 (0) Positive 11 (0) 13 (0) Indeterminate 41 (0) 192 (2) Note. Numbers in parentheses are the number of patients with adverse outcomes. * Treatment was considered a full course of anticoagulation therapy. US was generally performed with 5-MHz linear-array transducers by using compression as well as with color and duplex Doppler evaluation of the deep veins of the thigh and calf as described previously (18 23). The studies were performed by experienced sonographers, and all scans were reviewed by attending radiologists with special expertise in US. US scans of the calf were considered positive for DVT 26 Radiology April 1999 Gottlieb et al

TABLE 3 Summary of Parameters in the Three Patients with Adverse Outcomes Patient Finding at Calf US Symptom at Presentation Recent Surgery Other Risk Factors Outcome 1 Indeterminate Leg swelling Culdoplasty None No PE* 2 Indeterminate Shortness of Renal None No PE breath transplantation Previous DVT Yes Thigh DVT 3 Negative Leg swelling Total hip replacement * PE developed 5 days after baseline US. PE developed 5 months after baseline US. Popliteal vein thrombosis developed 2 weeks after baseline US. TABLE 4 Comparison of Risk of Adverse Outcomes between Patients Who Did and Those Who Did Not Recently Undergo Surgery Adverse Outcomes* Recent Surgery Yes No Yes (n 67) 3 (3.5) 84 (96) No (n 196) 0 (0) 196 (100) Note. Numbers in parentheses are percentages. * Adverse outcomes included two nonlethal PE and the development of a thigh DVT. Significantly different (P.028, Fisher exact test) from the rate of adverse outcomes in the nonsurgical group. if one or more of the paired deep calf veins (posterior tibial, peroneal, or anterior tibial) were not compressible and negative if all of the deep calf veins were visualized satisfactorily in their entirety without evidence of DVT. If none of those findings were present, the scan was classified as indeterminate. Indeterminate studies were those scans in which the full lengths of the paired deep veins were not satisfactorily visualized with color or duplex Doppler evaluation. Thrombi identified in the intramuscular or deep perforating veins were considered positive for DVT. US scans of the thigh were considered positive for DVT if one or more of the deep thigh veins (common femoral, superficial femoral, profunda femoral, and popliteal) were not compressible and negative if all deep thigh veins were compressible. If none of those findings were present, the scan was classified as indeterminate. Indeterminate thigh studies were those examinations in which there was a question about compressibility of a segment of a deep vein that could not be resolved with either color or duplex Doppler US. Statistical Evaluation We used the binomial distribution for proportions to estimate the CIs for the rates of adverse outcomes (PE, thigh DVT) in our study population. The Fisher exact and 2 tests (Instat; GraphPad Software, San Diego, Calif) were used to determine the statistical significance of associations of any of the parameters of our study population (eg, indications for US study, risk factors for DVT) with the presence of adverse outcomes. A result was considered statistically significant if the P value was less than.05. RESULTS Full Course of Anticoagulation Therapy Adverse outcomes were very rare, occurring in three of the 283 patients (1.1%; 95% CI 0.2%, 3.1%) (Table 2). Two of the three patients (0.7%; 95% CI 0.08%, 2.5%) had indeterminate US scans of the calf, did not receive anticoagulation therapy, and developed nonlethal PE documented with high-probability nuclear medicine lung scans. The third patient (0.4%; 95% CI 0%, 2.0%) had a negative calf US scan and corresponding conventional venogram for DVT, received anticoagulation therapy, and was documented with subsequent venography to have a thrombus in the popliteal vein. All three of these patients had recently undergone surgery (Table 3). Most of the 280 patients without adverse outcomes did not undergo additional studies to exclude a PE or lower extremity DVT. Forty-four (15.7%) of the 280 patients without adverse outcomes had low-probability or negative nuclear medicine lung scans for PE; two of these patients had negative pulmonary arteriograms and two had negative computed tomographic (CT) arteriograms for PE. Two patients (0.7%) had only follow-up CT arteriograms, which were negative for PE. Nine patients (3.2%) underwent conventional venography within 24 hours after indeterminate initial lower extremity US scans were obtained, with eight of these studies being negative and one positive for calf DVT in the peroneal and anterior tibial veins. The patient with calf DVT did not receive anticoagulation therapy. In 21 patients (7.5%), follow-up lower extremity US scans were negative for DVT in the thigh. No patient without a history of recent surgery (n 196) had an adverse outcome (0%; one-sided 97.5% CI 0%, 1.9%). US helped detect 17 patients in this group (8.7%) with isolated calf DVT. Seven of these patients (3.6%) received anticoagulation therapy. Recent surgery was the only parameter that showed correlation with an increased risk of adverse outcome (P.028) (Table 4). Adverse outcomes did not show correlation with other risk factors for DVT or PE, the results of the calf US study, or whether the patient received anticoagulation therapy. A history of recent surgery did not show correlation with any other risk factor for DVT or PE, which may have served to confound the results. DISCUSSION Clinical assessment has been documented to be an unreliable method for detecting DVT in the lower extremities (24 27). Some form of diagnostic testing has, therefore, been considered necessary to improve the accuracy of diagnosing DVT, with conventional venography being held as the standard of reference. US, due to its noninvasiveness and portability, has now replaced conventional venography as the primary imaging modality in the evaluation of patients clinically suspected of having DVT in the lower extremities. Conventional venography is considered more accurate than US in the diagnosis of calf DVT, particularly in asymptomatic patients (3 6). There is still much debate about the importance of isolated calf DVT as a substantial risk factor in the subsequent development of PE (7 16). We therefore considered it important to determine the clinical outcome of patients who presented to our US laboratory to exclude a lower extremity DVT in whom a negative US scan of the thigh was obtained. We evaluated in this group of patients the subsequent risk of developing a clinically important PE, considering if evaluation of the calf affected this outcome and whether any additional im- Volume 211 Number 1 Clinically Important Pulmonary Emboli: Does Calf Vein US Alter Outcomes? 27

aging was warranted to identify patients at risk of clinically important PE. A review of the literature identified two previous studies in which the clinical outcome of patients referred for lower extremity US to exclude a DVT was studied (28,29). No PE or thigh DVT occurred after a negative lower extremity US scan was obtained in the 431 patients followed-up in the study by Scarpa et al (28). The mean follow-up in the study by Scarpa et al was 8 months. The method of clinical follow-up, however, was not described in their study. PE occurred at an extremely low rate after a negative US scan was obtained in the 1,022 patients followed-up by Vaccaro et al (29); their method of follow-up was well characterized. Follow-up ranged from 8 to 33 months. In both studies, US evaluation was of the thigh only (unlike our protocol, in which both the thigh and calf were routinely evaluated). Subsequent thigh DVT occurred at a rate of 10.5% in the study by Vaccaro et al (29) after an initial negative result was obtained in patients with persistent lower extremity symptoms. This prompted Vaccaro et al to recommend follow-up US in patients in whom there remains a high clinical suspicion for lower extremity DVT. We found, as before, that the rate of clinically important PE following a negative thigh evaluation was very low, occurring in only two patients (0.7%) in our study group; both of these patients were treated successfully. These patients had indeterminate US evaluation of the calf. Only one patient (0.4%) was documented to have subsequently developed thigh DVT, despite receiving a full course of anticoagulation, after the initial thigh and calf US scans and conventional venogram were negative. Our rate of subsequent thigh DVT after an initially negative lower extremity US study is lower than that reported by Vaccaro et al (29). This may be related to the fact that we routinely evaluated the calf, whereas only the thigh was routinely evaluated by Vaccaro et al. Evaluation of the calf led to the detection of calf DVT in 24 patients (8.5%). Detection of these thrombi resulted in full anticoagulation regimens in 11 of these patients, which may have prevented propagation of these thrombi into the thigh. It is interesting, however, that no adverse outcomes (PE or thigh DVT) occurred in the 13 patients in whom US depicted calf DVT and who did not receive full anticoagulation. The retrospective nature of our study design prevented us from determining why some patients with calf DVT were treated and others were not. We could not evaluate how patient risk factors for anticoagulation, the extent of the thrombus, or the personal opinions of the physicians treating the patients entered into treatment decisions. There was no substantial difference between the rates of subsequent PE in our study and that of Vaccaro et al (29), despite the calf being routinely evaluated as part of our protocol and not evaluated as part of their protocol. Only patients who had recently undergone surgery developed PE. Similarly, the only patient who subsequently developed a thigh DVT also had a history of recent surgery. All of the adverse outcomes (PE and thigh DVT) occurred in postsurgical patients. Although adverse outcomes were still unusual in this group (3.5%; 95% CI 0.7%, 9.7%), the rate of adverse outcomes was significantly higher (P.028) than that in the nonsurgical group (0%; 97.5% one-sided CI 0.%, 1.9%). No other parameters, including all of the other risk factors for DVT or PE, indications for the study, type of treatment, or the results of the US study, were significantly associated with an increased risk of adverse outcomes. A history of recent surgery was not associated with any of the parameters described above, which may have served as confounders. Future larger prospective studies are necessary to confirm whether postsurgical patients are at increased risk of PE or propagated DVT into the thigh independent of the results of their lower extremity US examinations (negative, positive, or indeterminate). The results of our study must be considered preliminary in this regard because of the potential biases we may not be considering due to our retrospective study design and the relatively small number of adverse outcomes considered in our study population. Previous comparisons of lower extremity US findings in symptomatic and asymptomatic postsurgical patients have shown a difference in the accuracy of US for depicting both thigh and calf DVT. US has been shown to be less accurate for depicting both thigh and calf DVT in asymptomatic postsurgical patients than in symptomatic patients (3). This has led to many centers discontinuing screening US examinations as part of routine surveillance in postsurgical patients who are at increased risk of PE and lower extremity DVT due to their immobility (30). The results of our study support this philosophy, because these patients seem to be at higher risk for PE or propagated DVT into their thigh independent of the results of lower extremity US. Conversely, no adverse outcomes occurred in patients without a history of recent surgery with a negative thigh US scan. US of the calf in these patients resulted in only seven patients (3.6%) receiving full anticoagulation for isolated calf DVT. Ten patients (5.1%) did not receive full anticoagulation despite US depiction of an isolated calf DVT. No adverse outcomes occurred in either group, which suggests that it may not be necessary to administer full anticoagulation therapy to patients with isolated calf DVT to prevent the subsequent development of PE. Many physicians already follow this approach to treat isolated calf DVT. US of the calf is not warranted to screen patients at risk for subsequent PE if prophylaxis for this condition is not affected by the results of calf US. In the group of patients who did not recently undergo surgery, no adverse outcomes occurred in patients in whom US was considered negative for DVT in the thigh regardless of whether they received full anticoagulation. Certainly, additional evaluation of the calf (eg, conventional venography) is not indicated in these patients to identify individuals at risk for subsequent PE. One cannot justify an additional diagnostic study, such as venography, when the risk of complications from the test (thrombophlebitis in 1% 2% of patients) (31) exceeds the risk of developing the adverse outcomes one is attempting to prevent. We believe initial US evaluation of the calf to identify patients at risk for the subsequent development of PE is not necessary. US did not depict calf DVT in the patients who had adverse outcomes in our study group. The frequency of subsequent PE was exceedingly low in our population of patients with negative thigh studies, and PE did not occur in patients without a history of recent surgery. We also believe that follow-up US of patients with initially negative thigh studies should not be performed routinely because of the exceedingly low rate of subsequent PE or thigh DVT development in these patients. The decision to reevaluate the thigh with US should be tailored to individual patient risk factors for DVT and the presence or absence of local symptoms or physical findings. We recognize that calf US did depict DVT in 24 patients (8.5%) in our study population with negative thigh examinations and that 11 of these patients received anticoagulation therapy. One could argue that treatment of these patients prevented the development of subsequent PE. This argument, however, seems 28 Radiology April 1999 Gottlieb et al

unlikely given that no PE occurred in any of the patients (n 13) with documented calf DVT who did not receive anticoagulation therapy. Fifty-six (19.8%) of our patients received anticoagulation therapy, and 45 of these patients did not have a DVT. This could have reduced the rate of adverse outcomes in our study group. However, one patient with an adverse outcome (thigh DVT) received anticoagulation therapy and 227 patients who did not receive anticoagulation therapy did not develop an adverse outcome (13 with calf DVT). Routine evaluation of the deep venous system of the calf may be indicated if US depiction of DVT leads to more timely treatment or prevents chronic venous insufficiency due to damage to the venous valves. Patients with localized calf symptoms or physical findings most likely fall into this category because US can also depict alternative pathologic processes (Baker cyst, hematoma, neoplasm, abscess) in 12% of patients referred for suspected lower extremity DVT (29). We had a high rate of indeterminate calf US scans (82%) for DVT. This high rate most likely resulted from (a) use of the strict requirement of visualizing the full lengths of the deep calf veins with duplex or color Doppler US before calling an examination negative and (b) in many cases being unable to have patients dangle their legs over the side of the bed (using gravity to fill the deep calf veins) due to debilitating medical conditions. Despite our high indeterminate examination rate, very few adverse outcomes occurred, which further supports the concept of not routinely scanning the calf to identify patients at risk of clinically important PE. A potential limitation of our study is that autopsies were not performed in 27 of the 28 patients who died less than 6 months after initial lower extremity US. We believe that we are justified in considering the cause of death to be other than PE in these patients because a reasonable alternative was listed on the death certificate or in the medical record by the physician treating the patient. In summary, routine evaluation of the calf with US or other diagnostic imaging studies to identify patients at risk of developing clinically important PE is not indicated. Although follow-up thigh US is not routinely indicated in patients with initially negative thigh studies, it may be appropriate in select circumstances based on pretest risk factors for DVT or PE. A prospective study addressing whether calf US has a role in guiding therapy to control local symptoms from calf DVT or prevent the development of the postphlebitic syndrome would be useful. References 1. Gillum RR. Pulmonary embolism and thrombophlebitis in the United States, 1970 1985. Am Heart J 1987; 114:1262 1264. 2. Sandler DA, Martin JF. Autopsy proven pulmonary embolism in hospital patients: are we detecting enough deep vein thrombosis? J R Soc Med 1989; 82:203 205. 3. Rose SC, Zwiebel WJ, Murdock LE, et al. Insensitivity of color Doppler flow imaging for detection of acute calf deep venous thrombosis in asymptomatic postoperative patients. JVIR 1993; 4:111 117. 4. Cronan JJ. Venous thromboembolic disease: the role of US. Radiology 1993; 186: 619 630. 5. Rose SC, Zwiebel WJ, Nelson BD, et al. Symptomatic lower extremity deep venous thrombosis: accuracy, limitations, and role of color duplex flow imaging in diagnosis. Radiology 1990; 175:639 644. 6. Simons GR, Skibo LK, Polak JF, et al. Utility of leg ultrasonography in suspected symptomatic isolated calf deep venous thrombosis. Am J Med 1995; 99: 43 47. 7. Giachino A. Relationship between deepvein thrombosis in the calf and fatal pulmonary embolism. Can J Surg 1988; 31: 129 130. 8. Sevitt S, Gallagher N. Venous thrombosis and pulmonary embolism: a clinicopathological study in injured and burned patients. Br J Surg 1961; 48:475 489. 9. Philbrick JT, Becker DM. Calf deep venous thrombosis: a wolf in sheep s clothing? Arch Intern Med 1988; 148:2131 2138. 10. Huisman MV, Buller HR, Ten Cate JW, et al. Serial impedance plethysmography for suspected deep venous thrombosis in outpatients: the Amsterdam General Practitioner s Study. N Engl J Med 1986; 314:823 828. 11. Hull RD, Hirsh J, Carter CJ, et al. Diagnostic efficacy of impedance plethysmography for clinically suspected deep-vein thrombosis: a randomized trial. Ann Intern Med 1985; 102:21 28. 12. Hull RD, Hirsh J, Sackett DL, et al. Replacement of venography in suspected venous thrombosis by impedance plethysmography and I125-fibrinogen leg scanning: a less invasive approach. Ann Intern Med 1981; 94:12 15. 13. Lagerstedt CI, Olsson CG, Fagher BO, et al. Need for long-term anticoagulation treatment in symptomatic calf-vein thrombosis. Lancet 1985; 2:515 518. 14. Havig O. Deep vein thrombosis and pulmonary embolism: an autopsy study with multiple regression analysis of possible risk factors. Acta Chir Scand 1977; 478(suppl):42 47. 15. Rollins DL, Semrow CM, Friedell ML, et al. Progress in the diagnosis of deep venous thrombosis: the efficacy of real-time B-mode ultrasonic imaging. J Vasc Surg 1988; 7:638 641. 16. Kakkar VV, Howe CT, Flanc C, Clarke MB. Natural history of postoperative deepvein thrombosis. Lancet 1969; 2:230 232. 17. American Medical Association. International classification of diseases-9-cm. Vols 1 and 2. Chicago, Ill: American Medical Association, 1998. 18. Cronan JJ, Dorfman GS, Scola FH, et al. Deep venous thrombosis: US assessment using vein compression. Radiology 1987; 162:191 194. 19. Appleman PT, DeJong TE, Lampman LE. Deep venous thrombosis of the leg: US findings. Radiology 1987; 163:743 746. 20. Vogel P, Laing FC, Jeffrey RB Jr, et al. Deep venous thrombosis of the lower extremity: US evaluation. Radiology 1987; 163: 747 751. 21. Dauzat MM, Laroche JP, Charras C, et al. Real-time B-mode ultrasonography of better specificity in the noninvasive diagnosis of deep venous thrombosis. J Ultrasound Med 1986; 5:625 631. 22. Raghavendra BN, Horil SC, Hilton S, et al. Deep venous thrombosis: detection by probe compression of veins. J Ultrasound Med 1986; 5:89 95. 23. Cronan JJ, Dorfman GS, Grusmark J. Lower-extremity deep venous thrombosis: further experience with and refinements of US assessment. Radiology 1988; 168:101 107. 24. McLachlin J, Richards T, Paterson JC. An evaluation of clinical signs in the diagnosis of venous thrombosis. Arch Surg 1962; 85:738 744. 25. Nicolaides AN, Kakkar VV, Field ES, et al. The origin of deep vein thrombosis: a venographic study. Br J Radiol 1971; 44: 653 663. 26. Sandler DA, Martin JF, Duncan JS, et al. Diagnosis of deep-vein thrombosis: comparison of clinical evaluation, ultrasound, plethysmography, and venoscan with x- ray venogram. Lancet 1984; 2:716 719. 27. Wells PS, Hirsh J, Anderson DR, et al. Accuracy of clinical assessment of deepvein thrombosis. Lancet 1995; 345:1326 1330. 28. Scarpa MS, Messina LM, Villemure P, et al. Significance of a negative duplex scan in patients suspected of having acute deep venous thrombosis of the lower extremity. J Vasc Technol 1989; 13:224 226. 29. Vaccaro JP, Cronan JJ, Dorfman GS. Outcome analysis of patients with normal compression US examinations. Radiology 1990; 179:443 446. 30. Kearon C, Julian J, Newman TE, et al. Noninvasive diagnosis of deep venous thrombosis. Ann Intern Med 1998; 128: 663 676. 31. Foley WD, Middleton WD, Lawson TL, et al. Color Doppler ultrasound imaging of lower-extremity venous disease. AJR 1989; 152:371 376. Volume 211 Number 1 Clinically Important Pulmonary Emboli: Does Calf Vein US Alter Outcomes? 29