SEARCH

SEARCH BY CITATION

Keywords:

  • deep vein thrombosis;
  • public awareness;
  • pulmonary embolism;
  • venous thromboembolism

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Summary.  Background:  Although there have been attempts to raise public awareness about deep vein thrombosis (DVT), their influence on identifying confirmed cases is unknown.

Objective:  To determine the effect and its duration of a public awareness campaign about venous thromboembolism.

Patients/Methods:  A campaign to raise public awareness of DVT was conducted during one year in an urban population of approximately 100 000 (pop A). A comparison urban population of approximately 1 574 000 (pop B) was not exposed to this campaign. Patients symptomatic for DVT in both populations were referred by general practitioners for a standardized compression ultrasound (CUS) of the whole leg at no charge. Positive CUS examinations documented by photographs were analyzed by an independent adjudication committee blinded to the population. Pop A was followed for 8 months after the information campaign ended.

Results and Conclusions:  Symptomatic objectively confirmed DVT was found in 48 of 800 subjects tested in pop A and 226 of 2384 tested in pop B. The 1-year incidence of confirmed DVT (proximal and distal) was 46/100 000 (95% CI, 33–59) in A and 14/100 000 (95% CI, 12–16) in B (P < 0.001). The increase in pop A was due to distal DVT (36/100 000 vs. 5/100 000 in pop B, P < 0.001). The DVT rate for pop A in an 8-month follow-up period was 12/100 000, significantly lower than in the first 8 months of the study period (34/100 000/8 months) (P = 0.001). The public awareness campaign significantly increased the diagnosis of distal DVT. When the campaign ended, DVT rates returned to community baseline.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

Symptoms of deep vein thrombosis (DVT) are non-specific. Patients in whom objective tests confirm the diagnosis can have large thrombi but present with only minimal symptoms that, when disregarded, can result in dangerous consequences, including sudden death. However, the majority of DVT patients have a good prognosis. It is tempting to speculate that an increase in public awareness of DVT-related health threats and knowledge of its non-specific symptoms may in time reduce mortality and morbidity from venous thromboembolic disease (VTE). If true, increasing public awareness of symptoms and sequelae of DVT would be an important goal.

Such an increase in awareness is the aim of numerous governmental and non-governmental organizations, such as medical centers within the UK’s National Health Service, the US Centers for Disease Control and Prevention, and the Polish Foundation Against Thrombosis. However, we have found no studies reporting results of actions to increase public awareness of VTE or DVT. Therefore, we speculated that an increase in public awareness would result in an increase in the frequency of objectively confirmed DVT and conducted a controlled study. We believed it futile to do the study by randomizing within a population because control subjects would be exposed through conversation and other ways to the educational material. Hence, we chose two separate populations with similar demographic variables from different areas, exposed one to an educational campaign, and followed each.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The study was registered (http://www.controlled-trials.com/ISRCTN27831914). The study protocol was approved by the Ethics Committees of each participating study center. Each determined that informed consent for the educational component and for recording de-identified outcome results was not required. The educational intervention was conducted between December 2007 and December 2008 (the study period). Eligible subjects older than 20 years in two populations (pop) who were not taking any anticoagulation were eligible to be counted as DVT cases. Pop A included residents of a large city district comprised of 104 262 individuals belonging to 40 GP practices. Pop B comprised 449 GP practices covering 1 574 736 individuals from different cities.

Leaflets about DVT symptoms and sequelae of the disease were sent to addresses of houses and flats of all members of pop A four times during the year. Posters were hung that replicated the leaflet’s information, and the leaflets’ information was also published four times a year in local newspapers. Pop B was not exposed to specific education to increase public awareness of DVT-related symptoms and sequelae.

For all study patients presenting with DVT symptoms (pain, leg swelling, erythema, the patient reporting a palpable cord, the patient reporting superficial phlebitis), GPs referred the patient for urgent compression ultrasound (CUS) provided free of charge by trained specialist physicians at 16 investigator sites spread throughout the areas where the study was conducted. The GPs completed a form containing presenting symptoms for thromboembolism, Wells [1] DVT score, and relevant history and physical findings. D-dimer testing was not routinely done. Regardless of the estimated probability, all patients were to be referred for CUS. The specialists performing CUS made two reports of their findings. One was returned to the GP and one, along with a hard copy of the relevant CUS finding(s) and the GP’s completed form, was collected by independent study monitors. These were returned to the central study office in Warsaw for data tabulation and blinding of the CUS site prior to blinded adjudication. This was continued for pop A during the first 8 months of 2009 (follow-up period after the study period year) but no further education about VTE was presented during this follow-up period.

We performed CUS of the proximal and calf (peroneal, anterior tibial and posterior tibial) veins according to generally accepted principles [2] in the manner previously described [3]. The primary criterion for diagnosing DVT was loss of venous compressibility. The common and superficial femoral veins were examined with the patient in the supine position. For better visualization, the lower extremities were rotated externally. The veins were evaluated as distally as possible with the transducer held in both transverse and longitudinal position. To examine the popliteal vein, the patient was in either the supine or prone position with knees slightly flexed. When in the prone position, the legs were supported by the examiner’s fingers. Imaging of the calf was performed in the supine position with knees flexed. External rotation was employed if necessary, especially for examination of peroneal veins. The veins were identified above the ankles and followed superiorly as far as possible. A vein was considered positive for DVT when it was non compressible. A written description of the examination with photographic documentation of any lesion was analyzed by an independent adjudication committee blinded to the cluster from which the patient was derived.

For our sample size determination, we reasoned that a DVT incidence of 1/1000 population/year is frequently cited but we thought this might overestimate the incidence we could capture through GPs without our capturing night or weekend cases, so we reduced the expected rate by 25% for the comparison population, pop B, to 75/100 000. With the resources available, we thought we should show a 33% increase in identifying DVT cases in pop A, to 1/1000 (the oft-cited rate), if the educational effort were successful. Because pop B comparison group patients did not require the educational interventions, only close surveillance and payment for their CUS examinations, we planned to enroll a larger number to obtain a precise estimate. For a ratio of 15 to 1, control to intervention patients, using the above estimates, with 80% power to find a difference if one exists and a two-sided alpha of 0.05, we would require approximately 110 000 subjects in pop A and 1 660 000 in pop B, if our estimates were correct. Comparisons between quantitative variables were analyzed with the chi-square test. Calculations were carried out with stastica 7.0 software (StatSoft, Krakow, Poland). Although the study was sponsored by Sanofi-Aventis, it was entirely directed, analysed and written up by the otherwise independent investigator-authors.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

In pop A, comprised of somewhat younger persons (Table 1), 48 patients had an acute DVT detected during the 12-month study period, including 10 proximal DVTs and 38 distal DVTs. In pop B, 226 patients had DVT, including 146 proximal DVTs and 80 distal DVTs. The incidence of all, proximal and distal DVT events per 100 000 individuals in pop A was 46 (95% CI, 33.0–59.10), 9.6 (95% CI, 3.6–15.5) and 34.6 (95% CI, 24.9–48), respectively, and in pop B 14.4 (95% CI, 12.5–16.2), 9.3 (95% CI, 7.8–10.8) and 5.1 (95% CI, 4.0–6.2), respectively (Table 2).

Table 1.   Demographic characteristics of populations A and B
 Pop APop B
Number of people104 2621 574 736
Age 20–40 years old (%)4422
Age 41–60 years old (%)3640
Age 61–80 years old (%)1731
Age > 80 years old (%)37
Male (%)4742
Female (%)5358
Table 2.   Comparison of rates per 100 000 of all, proximal and distal DVT cases
Population N DVTProximalDistal
  1. Comparisons between populations were carried out using the chi-square test. Statistically significant differences between pop A and B in rates of all and distal DVT cases were observed (P < 0.0001 for both comparisons). There was no statistically significant difference between pop A and pop B in rates of proximal DVT.

A
 N 104 262481038
 N/100 00046.09.636.4
 95% CI33.0–59.13.6–15.524.9–48
B
 N 1 574 73622614680
 N/100 00014.49.35.1
 95% CI12.5–16.27.8–10.84.0–6.2
P(A vs. B) < 0.00010.92< 0.0001

In the 8-month follow-up period for pop A, symptomatic DVT was confirmed in 12 patients (four proximal and eight distal). During the corresponding first 8 months of the study period, 35 events of DVT were confirmed, four proximal and 31 distal. The rate of DVT in the follow-up period was 11.5/100 000/8 months and was significantly lower than in the corresponding 8 months of the study period, 33.6/100 000/8 months (P < 0.001). The estimated crude annual rate of DVT in the follow-up period in pop A was 17.4/100 000/12 months and was similar to the incidence of DVT observed in pop B, 14.4/100 000/year, during the study period.

Eight hundred CUS tests were performed in pop A (767 tests/100 000 individuals/year), of which 48 (6%) confirmed symptomatic DVT. In pop B, 2384 CUS tests were performed (151 tests per 100 000 individuals per year), of which 226 (9.5%) confirmed symptomatic DVT (P = 0.002). No symptomatic patient with a normal CUS test presented later during the study with confirmed venous thromboembolism.

The mean values (SD) of the Wells scores in patients with confirmation of DVT were similar, 3.5 (SD 1.9) in pop A and 3.2 (SD 1.5) in pop B (P = 0.12). Some characteristics of patients with confirmed DVT are listed in Table 3. No significant differences were observed regarding rates of confirmed pulmonary embolism (three patients in pop A (6%) and 18 patients in pop B (8%)), history of DVT, immobilization, cancer, respiratory failure, infection, rheumatic disease, history of stroke, paraplegia, varicose vein, obesity, dehydration, use of oral contraceptives, known thrombophilia, pregnancy, postpartum period, recent trauma and history of recent surgery in patients presenting with, or confirmed to have, DVT in pop A and B.

Table 3.   Characteristics of patients with confirmed DVT in populations A and B
 Patients with confirmed DVT in population APatients with confirmed DVT in population B
Number of confirmed events of DVT48226
Mean age (years)6259
Male (%)3041
Female (%)7059
Mean weight (kg)7884
Mean abdomen circumference (cm)99100
Symptoms of PE (%)68
History of VTE (%)1414
Immobilization (%)117
Cancer (%)64
Heart failure(%) class III/IV01
Respiratory failure (%)74
Infection (%)11
Rheumatic disease (%)33
History of stroke (%)31
Paraplegia (%)00
Varicose vein (%)3027
Obesity (%)1721
Dehydration (%)01
Contraception (%)13
Known thrombophilia (%)11
Pregnancy (%)01
Postpartum period (%)00
Trauma (%)31
History of recent surgery (%)32

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The results obtained confirmed that a public education campaign to raise awareness of clinical symptoms and risks related to DVT led to an increased incidence of DVT events confirmed by objective testing. Table 1 shows that pop A was younger than pop B and might have been expected to have a lower or similar incidence rate to pop B on that basis. However, the incidence of all DVT events and distal DVT events was higher in younger pop A, although clinical probability scoring was similar in both groups, suggesting it truly was education about symptoms and sequelae that led to the increased presentations and positive diagnoses of distal DVT. The lack of difference in proximal DVT diagnoses despite awareness education probably occurred because proximal vein symptoms are more apparent and worrying, which encouraged patients to visit their GPs. Perhaps the symptoms of distal DVT did not themselves concern the patient and it was the increased awareness of DVT-related complications (e.g. pulmonary embolism and post-thrombotic syndrome) that encouraged these patients to visit their GPs. This suggests that we have identified a not previously described factor affecting distal DVT incidence: prior patient awareness.

Our study populations enjoyed expedited CUS examinations compared with usual circumstances (i.e. CUS only after approval by an angiologist or vascular surgeon who also examined the patient). Expedited CUS examinations upon the GP’s suspicion were used in both treatment groups so cannot account for the results but may be an important part of increased case-finding after an education intervention.

Our study had limitations. Our incident DVT rates were lower than expected in both pop A and pop B. In retrospect, the incidence rate we projected for DVT was likely to be erroneous, considering our surveillance was for DVT rather than all VTE. A 2003 estimate of overall VTE incidence in the USA [4] showed 70–113/100 000 with 66% DVT; this would correspond to 46–76/100 000 DVT cases/year. A family practitioner database study from the United Kingdom published in 2007[5] found an overall VTE incidence of 74.5/100 000 but 40.3/100 000 for DVT alone; 19% of VTE cases in that study were associated with recent surgery (although that article’s title suggests all diagnoses were in outpatients, a close reading indicates that the medical records evaluated included complete records from hospitals, other referrals and all prescriptions). The DVT rates we found in Poland were not implausibly lower. We assume that recently postoperative patients with leg symptoms most likely visited their surgeons rather than primary care physicians and hence escaped surveillance in our catchment study methodology. Patients with DVT associated with various medical co-morbidities hospitalized for various reasons in medical or surgical wards are likely to have had DVT diagnosed there, rather than in our outpatient surveillance. A proportion of patients from both populations undoubtedly sought urgent medical assistance, leading to diagnosis in hospital emergency departments, particularly at night and weekends.

Our objective testing methodology might not be as sensitive as similar methods that additionally employ Doppler examination [6] but our approach was similarly applied to both populations and final results adjudicated blindly. Pop A’s younger age but higher incidence of DVT is counterintuitive except when considering the effect of the experimental intervention, education. While our algorithm for diagnosis omitted standardized D-dimer testing and hence was not optimal [7–9], it reflects pragmatic practice in our settings and that of many practices where CUS, rather than D-dimer, is used after clinical examination to confirm or refute a DVT diagnosis. Randomization would have been logistically impossible and although our control and intervention populations should be and appear similar except for mean age, randomization is the best method to minimize unforeseen differences in study groups. We cannot generalize regarding relative costs of the clinical approaches we tested because costs were not collected and vary widely in different societies where medical care, fees and salaries are organized in so very different ways.

Overall, while we were able to assess only the fraction of the annual number of incident symptomatic DVTs presenting to primary care physicians, there is no evidence to suggest that the proportions underestimated were not similar in each population. Our study embedded two controls: a contemporaneous control arm that was15 times larger of scattered GP practices well away from the site of the intervention, and a re-evaluation of pop A after the intervention concluded using identical methodology during an additional 8-month follow-up period. Our finding of a quite similar incidence of proximal DVT diagnosis but distinctly different distal DVT diagnosis rates in the study’s experimental arm during the intervention reinforces the legitimacy of our findings. The decay in the DVT diagnosis rate found over the subsequent 8 months in the control arm further reinforces the likelihood that our intervention effect was real. Regrettably, experiences of failure of educational preventive interventions (e.g., smoking cessation and dietary restriction for obesity) after a period of success are all too frequent in medicine.

We conclude that a public awareness campaign regarding the symptoms and sequelae of DVT resulted in an increase in total and distal DVT cases. This effect was transient and seemed to disappear when the education ended. If preserving protection against complications of untreated distal DVT is valued, public education, like public vaccinations, should probably be ongoing.

Addendum

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

W.Z. Tomkowski secured funding and led the study group. W.Z. Tomkowski, M. Dybowska, P. Kuca, P. Andziak, A. Jawień, D. Ziaja, G. Malek, M. Górska and B.L. Davidson participated in study design, data analysis and interpretation, and writing. M. Dybowska took charge of day-to-day study management under W.Z. Tomkowski’s direction. W.Z. Tomkowski, P. Kuca, P. Andziak, A. Jawień, D. Ziaja and G. Malek participated in data collection.

Acknowledgements

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The authors thank H. R. Buller for his help in analyzing the data and preparation of the manuscript. The study was sponsored by SanofiAventis. Guarantors: W.Z. Tomkowski and B. L Davidson.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References

The study was sponsored by Sanofi-Aventis. Sanofi-Aventis played no other role.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Addendum
  8. Acknowledgements
  9. Disclosure of Conflict of Interests
  10. References
  • 1
    Wells PS, Anderson DR. Diagnosis of deep vein thrombosis in the year 2000. Curr Opin Pulm Med 2000; 6: 30913.
  • 2
    Prandoni P, Cogo A, Bernardi E, Villalta S, Polistena P, Simioni P, Noventa F, Benedetti L, Girolami A. A simple ultrasound approach for detection of recurrent proximal-vein thrombosis. Circulation 1993; 88: 17305.
  • 3
    Tomkowski WZ, Davidson BL, Wisniewska J, Malek G, Kober J, Kuca P, Burakowska B, Oniszh K, Gallus A, Lensing AWA. Accuracy of compression ultrasound in screening for deep venous thrombosis in acutely ill mecical patients. Thromb Haemost 2007; 97: 1914.
  • 4
    White RH. The epidemiology of venous thromboembolism. Circulation 2003; 107(Suppl. 23): 148.
  • 5
    Huerta C, Johansson S, Wallander M-A, Rodriguez LAG. Risk factors and short-term mortality of venous thromboembolism diagnosed in the primary care setting in the United Kingdom. Arch Intern Med 2007; 167: 93543.
  • 6
    Stevens SM, Elliott CG, Chan KJ, Egger MJ, Ahmed KM. Withholding anticoagulation after negative result on duplex ultrasonography for suspected deep venous thrombosis. Ann Intern Med 2004; 140: 98591.
  • 7
    Büller HR, Ten Cate-Hoek AJ, Hoes AW, Joore MA, Moons KG, Oudega R, Prins MH, Stoffers HE, Toll DB, van der Velde EF, van Weert HC, AMUSE (Amsterdam Maastricht Utrecht Study on thromboEmbolism) Investigators. Safely ruling out deep venous thrombosis in primary care. Ann Intern Med 2009; 150: 22935.
  • 8
    Jaeschke R, Gajewski P, Bates SM, Douketis J, Solnica B, Crowther M, Leśniak W, Brozek JL, Schünemann HJ, Zawilska K, Tomkowski W, Undas A, Sznajd J, Nizankowski R, Bała MM, Guyatt G. 2009 evidence-based clinical practice guidelines for diagnosing a first episode of lower extremities deep vein thrombosis in ambulatory outpatients. Pol Arch Med Wewn 2009; 119: 5419.
  • 9
    Ten Cate-Hoek AJ, Toll DB, Buller HR, Hoes AW, Moons KG, Oudega R, Stoffers HE, van der Velde EF, van Veert HC, Prins MH, Joore MA. Cost-effectiveness of ruling out deep venous thrombosis in primary care vs. care as usual. J Thromb Haemost 2009; 7: 20429.