Family Practice Advance Access originally published online on January 7, 2005
Family Practice 2005 22(1):86-91; doi:10.1093/fampra/cmh718
Limited value of patient history and physical examination in diagnosing deep vein thrombosis in primary care
Julius Center for Health Sciences and Primary care, University Medical Center Utrecht, PO Box 85060, 3508 AB Utrecht, The Netherlands
Email: R.Oudega{at}knmg.nl
Received 16 June 2004; Accepted 27 September 2004.
Oudega R, Moons KGM and Hoes AW. Limited value of patient history and physical examination in diagnosing deep vein thrombosis in primary care. Family Practice 2004; 22: 86–91.
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Background. Properly ruling in or out deep vein thrombosis (DVT) is important because of the risk of developing pulmonary embolism if untreated and the risk of bleeding when treated with anticoagulants. In primary care, the diagnosis in suspected DVT is non-specific and the physician has to decide which patients to refer for further diagnostic work-up on the basis of patient history and physical examination alone.
Objective. To quantify which (combination of) items from patient history and physical examination contribute to the diagnosis of DVT in primary care.
Methods. A cross-sectional study design was chosen and the setting was all primary care physicians adherent to three local hospitals. 1325 consecutive patients consulting their primary care physician with symptoms suggestive of DVT were included.
Results. We studied 1325 patients with suspected DVT. The prevalence of thrombosis (assessed by means of compression ultrasonography) in these patients was 29%. Multivariate regression analysis of 17 candidate predictors resulted in nine independent predictors of DVT: male gender, duration of symptoms, malignancy, immobilization, leg trauma, pain when walking, oedema, calf circumference and dilated veins. The predictive value of the combination of the nine independent variables was low, reflected in the ROC area (as a combination of the sensitivity and specificity) of this model of 0.68. The low discriminative value was also exhibited in the numbers of DVT in the different risk categories. For example, in the low-risk group, the probability of DVT was still 15%. The diagnostic performance of patient history and physical examination was similar (and thus poor) in all clinically relevant subgroups.
Conclusion. Patient history and physical examination in patients suspected of DVT are of limited value for the primary care setting to identify patients with a low or high probability of DVT and thus in the decision to refer for further diagnostic work-up.
Keywords. Diagnosis, patient history, physical examination, primary health care, venous thrombosis.
|
Introduction |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Properly ruling in or out deep vein thrombosis (DVT) is important because untreated DVT patients may develop pulmonary embolism, whereas treatment with anticoagulants carries a risk of major bleeding.1–4 Particularly in primary care with its limited diagnostic tools, the diagnosis in suspected DVT is difficult because of a wide variety of non-thrombotic disorders which can mimic the clinical presence of DVT.5–9 The diagnostic tools available to the primary care physician are patient history and physical examination. Based on these, the physician has to decide which patients should be referred for further, more burdening and costly tests such as compression ultra sound (CUS) or venography.3,10,11
The difficulties in diagnosing DVT have been apparent for more than three decades, since the introduction of venography.6–12 Recently a few studies developed diagnostic prediction or decision rules using information from patient history and physical examination to stratify patients suspected of DVT into groups of low-, moderate- and high-risk of presence of DVT.13–15
These rules, however, were developed in the secondary care setting and not in typical primary care patients. Nevertheless, the diagnostic prediction rules are commonly applied both in secondary and primary care, even though it is well known that prediction rules derived from a secondary setting often show poor generalizability when applied in a primary care setting and vice versa.18–20 Indeed, various recent studies by our group and others showed that with the available rules, the primary care physician can not adequately identify patients with a low- or high-probability of having DVT.21–24 Thus, diagnostic rules developed in primary care are needed, since their performance in day-to-day general practice may be better. Also, no previous study on diagnosis of DVT investigated men and women separately, even though in women other (hormonal) risk factors play a role and therefore different diagnostic variables may be important.25,26 Similarly, in most studies and guidelines on diagnosis of DVT, patients with recurrent DVT are not included even though 15–20% of all patients suspected of DVT have a history of DVT.27–29 Hence, the question remains whether a diagnostic decision rule derived in primary care, based on symptoms and signs, is able to adequately estimate the presence or absence of DVT in relevant subgroups of primary care patients suspected of DVT. In this study we investigated 1325 primary care patients suspected of having DVT and used formal diagnostic probability modelling (multivariate logistic regression modelling) to quantify which combinations of items from patient history and physical examination contribute to the diagnosis of proximal DVT, and to what extent these combinations are able to discriminate between presence or absence of proximal DVT. As the data set was sufficiently large, we did several sub-analyses, for men and women and for patients with a first and recurrent DVT. We also analysed the data for the diagnosis proximal and/or isolated calf vein thrombosis.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Patients
In a region, adherent to three local hospitals in The Netherlands, we prospectively identified 1325 consecutive adult (over 18 years) patients who visited the primary care physician with clinically suspected DVT between January 1 1999 and July 31 2002. This suspicion was primarily based on the presence of a painful, swollen leg that existed not more than 30 days. After informed consent was given, the primary care physician systematically documented information on the patient's history and physical examination. In total there were 17 candidate diagnostic predictors. Patient history included, besides age and gender, previous DVT, family history of DVT, history of any malignancy, immobilization, recent surgery, leg trauma, pain when walking, ill-feeling and duration of the symptoms. Physical examination included calf tenderness, distension of collateral veins, oedema, swelling of the calf or the whole affected limb, and circumference of the calves. For women two additional candidate predictors were documented, i.e. use of oral contraception and hormonal substitution therapy. The study protocol was approved by the Medical Ethical Committee of the University Medical Center, Utrecht.
Reference standard
After the standardized work-up by the physician, all patients were referred to one of the three regional hospitals to determine the presence or absence of DVT using leg ultrasound. All patients underwent real-time B-mode compression ultrasonography (CUS) of the lower extremities with a standard 5 to 12-MHz linear-array transducer (system V GE/Sonotion).3 Colour Doppler imaging was used to identify venous vessels and venous flow patterns. In patients with a normal CUS measurement this test was repeated after 7 days. The physician who judged the result of the echo was blinded to the results of the patient history and physical examination.
Proximal DVT was considered present if one of the two CUS measurements was abnormal. An abnormal CUS was defined if the iliac veins, the common femoral vein, the superficial femoral vein or the popliteal vein until the trifurcation were not completely compressible. Obstruction of the iliac veins was tested with help of colour Doppler.3,30–33 Isolated calf vein thrombosis was considered present if the anterior tibial, the posterior tibial or the peroneal veins could not be compressed completely.
Data analysis
In the total study population, we first quantified the association between each finding from patient history and physical examination and the presence or absence of proximal DVT using univariate analysis. As the number of patients in our study was large enough, we subsequently included all 17 candidate predictors together in a multivariate logistic regression model. Variables with a multivariate P-value <0.10, according to the likelihood ratio test, were retained in the final diagnostic model and were considered as independent predictors of the presence or absence of DVT. Calibration or reliability of the final diagnostic model was evaluated by comparing the predicted probability versus observed frequency of DVT for deciles of patients and tested using the Hosmer & Lemeshow test.34 The ability of the model to discriminate between patients with and without DVT was estimated by the area under the Receiver Operating Characteristic curve (ROC area) of the model.35 The ROC area can range from 0.5 (no discrimination, like a coin flip) to 1.0 (perfect discrimination).34,35 Generally a value over 0.7 is interpreted as reasonable and over 0.8 as good.36 Since the ROC area reflects the overall value of the model, it does not directly indicate its clinical value.37,38 Hence, we estimated the absolute number of correctly predicted patients with and without DVT across different categories of the model's predicted probability.
This entire analytical approach was then repeated for men and women, and for patients with a first or recurrent DVT, separately. The diagnostic outcome of the above analysis was proximal DVT, but we also did the analysis for the outcome ‘proximal DVT or isolated calf vein thrombosis’.
One hundred and twenty-seven subjects had missing values for one or more predictors. The average proportion of missing values per predictor was 4%. Incompleteness of data is seldom a random phenomenon. Therefore, deleting subjects with a missing value commonly leads to biased results and certainly to a loss of power.39–41 To decrease bias and increase statistical efficiency, it is better to impute missing values rather than doing a complete case analysis.39–41 Accordingly, we imputed our missing data using the linear regression method available in SPSS (version 11.0).
|
Results |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Total population
The mean age of the 1325 patients was 61 years and 22% were male. The overall prevalence of DVT was 29%. Table 1 shows the distribution of all 17 candidate predictors separately across patients with and without DVT. Twelve were associated with DVT, i.e. gender, malignancy, immobilization, surgery, leg trauma, pain when walking, duration of symptoms, dilated veins, oedema, swelling of the calf, swelling of the whole leg and difference in calf circumference.
|
Multivariate regression analysis (Table 2) showed that of the 17 candidate predictors, nine independently contributed to the prediction of DVT. The model including these nine predictors showed good calibration (Hosmer & Lemeshow test P-value of 0.55). The ROC area of this model however was 0.68 [95% confidence interval (CI) 0.65–0.71]. Following previous studies28,29,42 we also defined three probability or risk groups based on the model's estimated probability: low risk (probability <0.20), medium risk (0.20–0.80) and high risk (
0.80). Using these categories, more than a quarter of all patients (28%) fell in the lowest risk group, of which 15% still had DVT (Table 3). In the high-risk group all patients had DVT, but this group included only three patients (0.2%). The largest group of patients (72%) was included in the moderate risk category, with a DVT prevalence of 35%.
|
|
Subgroup analyses
In the 462 male patients the DVT prevalence was 37%, compared to 25% in the 863 female patients. For women two additional predictors were investigated: oral contraception use and hormonal supplementation therapy. Twelve percent of the women with DVT used oral contraception, whereas this was nine percent among women without DVT [odds ratio (OR) in multivariate analysis = 1.3; 95% CI 0.8–2.1]. Hormonal substitution therapy showed an OR in multivariate analysis of 1.6 (95% CI 0.8–3.3). Men and women had five independent predictors in common: leg trauma, ill feeling, oedema, calf difference and dilated veins. Additional independent predictors specific for men were age, surgery and pain when walking, while in women malignancy, oral contraception and duration of symptoms were additional predictors. The ROC area of the final diagnostic model for both men (0.69) and women (0.67) was comparable to the area found for the final model of the total population (Table 4).
|
Analysis of the 1118 patients without a previous DVT showed a DVT prevalence of 30%, while this was 26% among those with recurrent DVT and 29% in the total study population. Among patients with a first DVT, the final diagnostic model showed a ROC area of 0.69, whereas among those with recurrent DVT this was 0.68 (Table 4).
When the outcome was defined as proximal DVT and/or isolated calf vein thrombosis, we found an expected higher prevalence: 36%. The ROC area of the final model to predict the presence or absence of this combined diagnostic outcome was 0.64 (Table 4).
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
We investigated a large group of patients suspected of DVT in primary care to quantify the value of patient history and physical examination in diagnosing of DVT using formal multivariate probability modelling. We found nine independent predictors of DVT, but in combination these showed a very low ability to discriminate between the presence and absence of DVT. Also in relevant subgroups according to gender and first or recurrent DVT, and when the outcome was defined as proximal and/or distal DVT, the predictive value of patient history and physical examination was low.
To appreciate the present results various issues should be addressed. First, the diagnosis of DVT was made using repeated leg ultrasound and not by venography, the ‘gold standard’. Venography is invasive, expensive and has a chance of allergic reactions to contrast and iatrogenic thrombosis. The validity of replacing the gold standard venography by serial compression ultrasonography has been established.8,32 Because of the lower sensitivity of ultrasound to detect pelvic thrombi and distal DVT, some cases with DVT could have been missed and classified as absent. On the other hand, even when performed by experienced radiologists, 5%–12% of venograms are inadequate.3 Moreover, we explicitly chose for leg ultrasound as reference test as it also complies to current practice (venography is currently hardly practiced) and since this reference test was also applied in most recent studies on diagnosis of DVT. Second, some patients had missing values, which were imputed. Although this may seem controversial, it is widely acknowledged that such imputations lead to more valid and precise results than simply deleting subjects with missing data.40,41 Third, the clinical domain of this study was the primary care setting where first presentation of DVT occurs and the clinical data were collected by GPs. This certainly enhances the applicability of our findings in primary care, but the results should not be extrapolated to secondary care.
Our study has several strengths. First, all consecutive patients suspected of DVT in a defined region were investigated, including those with recurrent DVT. Second, we used multivariate regression modelling to quantify the discriminative value of combinations of items from patient history and physical examination. Last but not least to our knowledge it is the largest study on primary care patients exclusively, which enabled various subgroup analyses. In this large study, investigated solely in the domain of primary care, patient history and physical examination showed poor predictive value to rule in or out DVT, also in all relevant subgroups.
In 1997 Wells et al.13 published a diagnostic rule using only patient history and physical examination characteristics to categorise secondary care patients in low, medium or high probability of DVT presence. In their study the low risk category included 44% of all patients suspected of DVT of which only 3% had DVT. In our study, similarly using a diagnostic rule also including only patient history and physical examination characteristics, we found in the low risk category a much lower percentage of patients (28%) of which still 15% had DVT. The latter was unacceptably high. Furthermore, in the high risk category, Wells et al. included 12% of all patients of which 75% indeed had DVT, whereas in our study only 0.2% of all patients was included in the high risk category. This illustrates the differences between patients in the primary and secondary care setting. This was also exhibited by differences in patient characteristics between our study and the Wells study. For example, in patients with presence of DVT, the frequency of malignancy in our study was 12% whereas in the Wells study it was 39%. Also, among patients with DVT, in our study 25% underwent surgery while this was 50% in Wells study. In addition, various recent studies tested the Wells rule in a primary care setting and showed that its applicability is limited.22–24
Several recent studies executed in secondary care evaluated a diagnostic decision rule including clinical features together with the D-dimer test result.43–45 The concentration of D-dimer, a degradation product of cross-linked fibrin, is always raised in the event of an acute DVT. These studies43–45 showed that in secondary care, anticoagulant therapy can be safely withheld in patients with low or medium risk of DVT—based on clinical features—together with a negative D-dimer test result. Based on these studies, one may infer that also in primary care it could be safe not to refer patients with a low or medium risk of DVT and a negative D-dimer test. Presently, however, there is yet no evidence for this conclusion. Accordingly, future studies carried out in primary care need to be done to prove the safety of a management strategy based on risk assessment using patient history and physical examination in combination with (rapid) D-dimer testing.
In conclusion, in primary care, patient history and physical examination alone are not sufficient to include or exclude DVT, and also not conclusive to properly select patients for referral for further work-up in secondary care. Primary care physicians need additional tools for the decision which patients to refer, such as a D-dimer test. Given the knowledge that diagnostic rules for ruling in or out DVT developed in secondary care have limited value in primary care, future studies should aim to quantify the added value of these tools to the patient history and physical examination in a primary care population.
|
Declaration |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
Funding: This work was supported by a grant from the Health Care Research Foundation IJsselmond in the Netherlands.
Ethical approval: Medical Ethical Committee of the University Medical Center, Utrecht.
Conflicts of interest: none.
|
References |
|---|
|
TopAbstractIntroductionMethodsResultsDiscussionDeclarationReferences |
|---|
1 Schulman S, Rhedin AS, Lindmarker P, Carlsson A, Larfars G, Nicol P et al. A comparison of six weeks with six months of oral anticoagulant therapy after a first episode of venous thromboembolism. Duration of Anticoagulation Trial Study Group. N Engl J Med 1995; 332: 1661–1666.
2 Lensing AW. Diagnosis of venous thrombosis. In Colman RW, Hirsch J, Marder VJ, Salzman EW (eds). Hemostasis and thrombosis: basic principles and clinical practice (3rd edn). Philadelphia: Lippincot; 1995, 1297–1322.
3 Fraser JD, Anderson DR. Deep venous thrombosis: recent advances and optimal investigation with US. Radiology 1999; 211: 9–24.
4 Hirsh J, Dalen JE, Anderson DR, Poller L, Bussey H, Ansell J et al. Oral anticoagulants: mechanism of action, clinical effectiveness, and optimal therapeutic range. Chest 1998; 114: 445S–614S.[ISI][Medline]
5 Cranley JJ, Canos AJ, Sull WJ. The diagnosis of deep venous thrombosis. Fallibility of clinical symptoms and signs. Arch Surg 1976; 111: 34–40.[Abstract]
6 Haeger K. Problems of acute deep venous thrombosis. I The interpretation of signs and symptoms. Angiology 1969; 20: 219–242.
7 Buller HR, Lensing AW, Hirsh J, ten Cate JW. Deep vein thrombosis: new non-invasive diagnostic tests. Thromb Haemost 1991; 66: 133–140.[ISI][Medline]
8 Lensing AW, Prandoni P, Brandjes D, Huisman PM, Vigo M, Tomasella G et al. Detection of deep-vein thrombosis by real-time B-mode ultrasonography. N Engl J Med J 1989; 320: 342–347.
9 Anand SS, Wells PS, Hunt D, Brill-Edwards P, Cook D, Ginsberg JS. Does this patient have deep vein thrombosis? J Am Med Assoc 1998; 279: 1094–1103.
10 Hirsh J, Lee AY. How we diagnose and treat deep vein thrombosis. Blood 2002; 99: 3102–3112.
11 Michiels JJ, Kasbergen HA, Oudega R, Graaf Fvd, Maeseneer MD, Planken Mvd et al. Exclusion and diagnosis of deep vein thrombosis in outpatients by sequential noninvasive tools. Int Angiol 2002; 21: 9–19.[ISI][Medline]
12 McLachlin J, Richards T, Paterson JC. An evaluation of clinical signs in the diagnosis of deep vein thrombosis. Arch Surg 1962; 85: 733–777.
13 Wells PS, Anderson DR, Bormanis J, Guy F, Mitchell M, Gray L et al. Value of assessment of pretest probability of deep-vein thrombosis in clinical management. Lancet 1997; 350: 1795–1803.[CrossRef][ISI][Medline]
14 Landefeld CS, McGuire E, Cohen AM. Clinical findings associated with acute proximal deep vein thrombosis: a basis for quantifying clinical judgment. Am J Med 1990; 88: 382–390.[CrossRef][ISI][Medline]
15 Kahn SR, Joseph L, Abenhaim L, Leclerc JR. Clinical prediction of deep vein thrombosis in patients with leg symptoms. Thromb Haemost 1999; 81: 353–360.[ISI][Medline]
16 Hull RD, Stein PD, Ghali WA, Cornuz J. Diagnostic Algorithms for Deep Vein Thrombosis: Work in Progress. Am J Med 2002; 113: 687–695.[CrossRef][ISI][Medline]
17 Michiels JJ, Schroyens W, Maeseneer MD, Karsbergen HAA, Oudega R. The Rehabilitation of Clinical Assessment in the Diagnosis of Deep Vein Thrombosis. Seminars in Vascular Medicine 2002; 2: 1–5.[CrossRef][Medline]
18 Wasson JH, Sox HC, Neff RK, Goldman L. Clinical prediction rules. Applications and methodological standards. N Engl J Med 1985; 313: 793–802.[Abstract]
19 Justice AC, Covinsky KE, Berlin JA. Assessing the generalizability of prognostic information. Ann Intern Med 1999; 130: 515–539.
20 McGinn TG, Guyatt GH, Wyer PC, Naylor CD, Stiell IG, Richardson WS. Users' guides to the medical literature: XXII: how to use articles about clinical decision rules. Evidence-Based Medicine Working Group. J Am Med Assoc 2000; 284: 79–84.
21 Cornuz J, Ghali WA, Hayoz D, Stoianov R, Depairon M, Yersin B. Clinical prediction of deep venous thrombosis using two risk assessment methods in combination with rapid quantitative D-dimer testing. Am J Med 2002; 112: 198–203.[CrossRef][ISI][Medline]
22 Schutgens RE, Ackermark P, Haas FJ, Nieuwenhuis HK, Peltenburg HG, Pijlman AH et al. Combination of a normal D-dimer concentration and a non-high pretest clinical probability score is a safe strategy to exclude deep venous thrombosis. Circulation 2003; 107: 593–598.
23 Tick LW, Ton E, van Voorthuizen T, Hovens MM, Leeuwenburgh I, Lobatto S et al. Practical diagnostic management of patients with clinically suspected deep vein thrombosis by clinical probability test, compression ultrasonography, and D-dimer test. Am J Med 2002; 113: 630–635.[CrossRef][ISI][Medline]
24 Oudega R, Hoes AW, Moons KG. Validation of the Wells' rule to exclude deep vein thrombosis in primary care. Submitted 2004.
25 Jick H, Jick SS, Gurewich V, Myers MW, Vasilakis C. Risk of idiopathic cardiovascular death and nonfatal venous thromboembolism in women using oral contraceptives with differing progestagen components. Lancet 1995; 346: 1589–1682.[CrossRef][ISI][Medline]
26 Rosendaal FR, Helmerhorst FM, Vandenbroucke JP. Oral contraceptives, hormone replacement therapy and thrombosis. Thromb Haemost 2001; 86: 112–135.[ISI][Medline]
27 Wells PS, Rodger M, Forgie M, Anderson DR, Kovacs M, Florack P et al. A randomised trial in patients with suspected DVT comparing a D-dimer/clinical probability strategy to clinical probability, prior to ultrasound imaging. D-dimer safely reduces the need for diagnostic imaging. Thromb Haemost 2001; 86: SOC41.
28 Cornuz J, Ghali WA, Hayoz D, Stoianov R, Depairon M, Yersin B. Clinical prediction of deep venous thrombosis using two risk assessment methods in combination with rapid quantitative D-dimer testing. Am J Med 2002; 112: 198–203.[CrossRef][ISI][Medline]
29 Perrier A, Desmarais S, Miron MJ, de Moerloose P, Lepage R, Slosman D et al. Non-invasive diagnosis of venous thromboembolism in outpatients. Lancet 1999; 353: 190–195.[CrossRef][ISI][Medline]
30 Hirsh J, Hoak J. Management of deep vein thrombosis and pulmonary embolism. A statement for healthcare professionals. Council on Thrombosis (in consultation with the Council on Cardiovascular Radiology), American Heart Association. Circulation 1996; 93: 2212–2257.
31 Heijboer H, Buller HR, Lensing AW, Turpie AG, Colly LP, ten Cate JW. A comparison of real-time compression ultrasonography with impedance plethysmography for the diagnosis of deep-vein thrombosis in symptomatic outpatients. N Engl J Med 1993; 329: 1365–1374.
32 Cogo A, Lensing AW, Koopman MM, Piovella F, Siragusa S, Wells PS et al. Compression ultrasonography for diagnostic management of patients with clinically suspected deep vein thrombosis: prospective cohort study. Br Med J 1998; 316: 17–20.
33 Sluzewski M, Koopman MM, Schuur KH, van Vroonhoven TJ, Ruijs JH. Influence of negative ultrasound findings on the management of in- and outpatients with suspected deep-vein thrombosis. Eur J Radiol 1991; 13: 174–181.[CrossRef][ISI][Medline]
34 Hosmer DW, Lemeshow S. Confidence interval estimates of an index of quality performance based on logistic regression models. Stat Med 1995; 14: 2161–2233.[ISI][Medline]
35 Hanley JA, McNeil BJ. The meaning and use of the area under a receiver operating characteristic (ROC) curve. Radiology 1982; 143: 29–36.
36 Weinstein MC. Clinical decision Analysis. Philadelphia: WB Saunders; 1980.
37 Moons KG, van Es GA, Michel BC, Buller HR, Habbema JD, Grobbee DE. Redundancy of single diagnostic test evaluation. Epidemiology 1999; 10: 276–357.[CrossRef][ISI][Medline]
38 Moons KG, Stijnen T, Michel BC, Buller HR, van Es GA, Grobbee DE et al. Application of treatment thresholds to diagnostic-test evaluation: an alternative to the comparison of areas under receiver operating characteristic curves. Med Decis Making 1997; 17: 447–501.
39 Harrell FE Jr, Lee KL, Mark DB. Multivariable prognostic models: issues in developing models, evaluating assumptions and adequacy, and measuring and reducing errors. Stat Med 1996; 15: 361–448.[CrossRef][ISI][Medline]
40 Greenland S, Finkle WD. A critical look at methods for handling missing covariates in epidemiologic regression analyses. Am J Epidemiol 1995; 142: 1255–1319.
41 Little RJA. Regression with missing X's: a review. J Am Stat Assoc 1992; 87: 1227–1264.[CrossRef][ISI]
42 Miron MJ, Perrier A, Bounameaux H. Clinical assessment of suspected deep vein thrombosis: comparison between a score and empirical assessment. J Intern Med 2000; 247: 249–303.[CrossRef][ISI][Medline]
43 Kearon C, Ginsberg JS, Douketis J, Crowther M, Brill-Edwards P, Weitz Jl et al. Management of suspected deep venous thrombosis in outpatients by using clinical assessment and D-dimer testing. Ann Intern Med 2001; 135: 108–111.
44 Anderson DR, Kovacs MJ, Kovacs G, Stiell I, Mitchell M, Khoury V et al. Combined use of clinical assessment and D-dimer to improve the management of patients presenting to the emergency department with suspected deep vein thrombosis (the EDITED Study). JTH 2003; 1: 645–651.
45 Schutgens R, Ackermark P, Haas FJ, Nieuwenhuis HK, Peltenburg HG, Pijlman AH et al. Combination of a normal D-dimer concentration and a non-high pretest clinical probability score is a safe strategy to exclude deep vein thrombosis. Circulation 2003; 107: 593–597.
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  Thromboembolism after pneumonectomy for malignancy: An independent marker of poor outcome. J. Thorac. Cardiovasc. Surg., March 1, 2006; 131(3): 711 - 718.
|
|
|
|
 |
|
 R. Oudega, A. W. Hoes, and K. G.M. Moons The Wells Rule Does Not Adequately Rule Out Deep Venous Thrombosis in Primary Care Patients Ann Intern Med, July 19, 2005; 143(2): 100 - 107. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Goodacre, A. J. Sutton, and F. C. Sampson Meta-Analysis: The Value of Clinical Assessment in the Diagnosis of Deep Venous Thrombosis Ann Intern Med, July 19, 2005; 143(2): 129 - 139. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||