Inflammation in deep vein thrombosis and the development of post-thrombotic syndrome: a prospective study


Edith Roumen-Klappe, Department of General Internal Medicine, UMC St Radboud Nijmegen, PO-box 9101, 6500 HB Nijmegen, the Netherlands.
Tel.: +31 24 3618819; fax: +31 24 3541734.


Summary. Background: The aim of this study was to investigate whether inflammatory markers (interleukin-6 [IL-6] and C-reactive protein [CRP]) in the acute phase of deep vein thrombosis (DVT) are associated with elevated venous outflow resistance (VOR), thrombosis score (TS), reflux and the development of clinical post-thrombotic syndrome (PTS). Methods: In 110 patients with a first DVT, plasma concentrations of IL-6 and CRP were determined on the day of admission. VOR, TS and reflux were measured 7 days, 1 and 3 months after diagnosis. After 1 year patients were evaluated for PTS using the Clinical, Etiologic, Anatomic and Pathophysiologic (CEAP) classification and Villalta scale. Results: Median levels of IL-6 and CRP were 7 pg mL−1 and 21 mg L−1, respectively. After 3 months, VOR was elevated in 33 patients (30%), TS in 33 (30%) and reflux in 57 (52%). Incidence of PTS was 36.7% using CEAP ≥ 3 and 35.4% using Villalta-scale ≥ 5. Elevated levels of IL-6 and CRP were related to higher outcomes of VOR after 3 months [relative risks (RR) 2.4 (95% CI 1.5–3.9) and 1.4 (1.1–3.3), respectively] and for IL-6 to TS [1.5 (1.1–2.1)]. For reflux no relation was found. After 90 days, elevated outcomes of VOR, TS and reflux were related to PTS after 1 year. The association of IL-6 and CRP with PTS was weak using the CEAP classification with a RR of 1.2 (0.7–2.2) and 1.8 (0.9–3.3) and absent according to the Villalta scale 0.6 (0.2–1.4) and 1.2 (0.6–2.5), respectively. Conclusion: The results of this study suggest that inflammation might play a role in incomplete thrombus clearance, venous outflow obstruction and the development of PTS after 1 year.


Post-thrombotic syndrome (PTS) is a chronic and frequent complication of deep vein thrombosis (DVT). Symptoms are persistent or intermittent pain, heaviness, swelling, itching, tingling or cramping and are typically aggravated by standing or walking and tend to improve with rest and recumbency. Typical clinical signs of PTS include edema, venous ectasia, hyperpigmentation, eczema, varicose collateral veins and, in severe cases, lipodermatosclerosis and venous ulceration [1]. Despite the use of elastic stockings, PTS occurs in up to 20–50% of patients with DVT, impairs the quality of life [2] and has a substantial impact on health care costs [3,4].

The clinical manifestations of PTS are probably as a result of venous hypertension, which occurs as a consequence of venous valvular incompetence, persistent deep vein outflow obstruction, or both. In a previous study, we have shown that during the follow-up of DVT the occurrence of PTS after 1 year was best predicted by an increased venous outflow resistance (VOR), residual thrombus [measured as thrombosis score (TS)] and presence of reflux at 3 months [5].

The destruction of venous valves and persistent outflow obstruction might be caused by the volume effect of the thrombus, but also by an inflammatory process which accompanies thrombosis. Studies in animals showed that inflammation in DVT involves both the venous vascular wall and the perivascular area [6,7]. In patients with DVT, we have demonstrated previously that inflammatory markers such as interleukin-6 (IL-6) and C-reactive protein (CRP) are elevated as a result rather than a cause of venous thrombosis[8]. To our knowledge, no data exist examining the role of inflammation in the pathophysiology of PTS.

The aim of this study was to investigate whether increased levels of the inflammatory markers IL-6 and CRP in the acute phase of DVT are associated with elevated VOR, higher TS, reflux and with the subsequent development of clinical PTS.


Patients and study design

Between 2002 and 2005 consecutive outpatients with symptomatic DVT, confirmed by compression ultrasound, were considered for inclusion in the Post Thrombosis Study. Both patients with proximal thrombosis, involving the popliteal vein or above, and patients with distal thrombosis limited to the calf were included. Exclusion criteria were previous DVT, active malignancy, life expectancy less than 2 years, underlying infection, arterial insufficiency (Fontaine ≥ 2b), use of anti-inflammatory medication, previous venous insufficiency, surgery, trauma or pregnancy in the 4 weeks prior to the DVT and the use of low-molecular weight heparin (LMWH) or oral anticoagulation before blood sampling. The study was approved by the local ethics committee and all patients gave informed consent.

All participants were treated with LMWH (dose based on body weight) and vitamin K antagonists [international normalized ratio (INR) 2–3) for at least 3 months. They were instructed to wear sized-to-fit elastic stockings for at least 2 years. The stockings prescribed were class III (40 mmHg) thigh-length in case of proximal DVT for the first 3 months followed by knee-length stockings, or knee-length stockings only in case of distal DVT. For patients with arterial insufficiency or above 75 years of age, class II (30 mmHg) stockings were prescribed. Compliance was assessed by asking the patients. Almost all patients were treated in the outpatient clinic.

Outcome measures of venous non-invasive examinations

Follow-up visits and non-invasive examinations were planned after 7, 30 and 90 days.

VOR was measured in both extremities using strain-gauge plethysmography with the patient in a supine position with pneumatic cuffs around the thighs and strain gauges around the calves [9,10]. After inflation of the cuffs the venous volume increases gradually. When venous pressure equals the effective congestive pressure and maximum volume was achieved, the cuff pressure was released resulting in a volume decrease measured by plethysmography. At the point 0.5 s after pressure release, the tangent to volume record was drawn and the slope converted into a flow rate, which is the venous emptying rate (VER). This procedure is performed with five different cuff pressures. Subsequently, the corresponding VER values are plotted against the effective cuff pressures. The slope of the line through the point gives an angle β. Using analogy with Ohm’s law, the VOR is calculated as 1/angle β. The VOR value is expressed in resistance units: (mmHg × min)/%. A VOR exceeding 0.8 mmHg × min/% was considered abnormal [9,10]. The non-thrombotic leg was used as a control.

Duplex examinations were performed by two experienced vascular technicians using a Toshiba (Toshiba Medical Systems Corporation, Tokyo, Japan) SSA 270A scanner with a 3.75 or 5 MHz probe in a low-flow setting. The examination was performed with the patient in a reverse Trendelenburg position at 45 degrees, the knees flexed and the feet resting on a foot support. Sixteen vein segments were examined: the common femoral vein, the femoral vein (proximal and middle), the great saphenous vein (proximal, middle and distal), the popliteal vein, the short saphenous vein, the posterior and anterior tibial veins, the peroneal and gastrocnemial veins. In the calf, each artery is accompanied by two veins; these were numbered: one being the most superficial and two the deeper vein. The distal femoral vein is often not compressible because of the anatomical position in Hunter’s canal. After the third month, the peroneal, gastrocnemial and tibial anterior veins were only examined if thrombus was present at the initial examination, as these measurements were time-consuming and yielded little data. The results of the remaining 10-vein segments were used for analysis. The compressibility was assessed in the transverse plane with gentle pressure applied to the overlying skin using the duplex-probe. In the longitudinal plane, the presence of the Doppler signal is usually present; in the distal veins, a flow signal was considered present after gently squeezing the extremity below the level of measurement. The definition of the thrombosis score (TS) was based on the description of extent of thrombus and site involvement by Porter [11]. A vein segment was considered occluded when the vein segment was non-compressible and no Doppler signal was found (TS = 2) and non-occluded if it was not completely compressible but flow was still present (TS = 1). A fully patent vein segment showed complete compressibility and flow (TS = 0). The TS values were calculated by counting the TS values of the 10-vein segments [12,13].

Reflux was measured in the longitudinal plane and provoked in the proximal veins by the Valsalva maneuver and in the distal veins by manual compression with sudden release. Pathological reflux was defined as a reversed flow duration of more than 1 s in the proximal veins and more than 0.5 s in the distal veins [9,13,14].

Determination of PTS

After 1 year, patients were evaluated by an experienced dermatologist who was unaware of the results of the venous examination. The severity of PTS was scored according to the clinical score (range 0–6) of the revised Clinical, Etiologic, Anatomic and Pathophysiologic (CEAP) classification [15] and the Villalta scale [16]. In the CEAP classification, patients with class 0 represent no visible or palpable signs of venous disease symptoms of PTS; class 1 telangiectasias, reticular veins or malleolar flare; class 2 varicose disease; class 3 edema without skin changes; class 4 skin changes ascribed to venous disease (pigmentation, lipodermatosclerosis); class 5 skin changes with healed ulcer and class 6 skin changes with active ulceration. PTS was defined as a CEAP classification of ≥ 3 or more. The Villalta scale, also called the Prandoni scale, grades the severity from 0 to 3 of five symptoms (pain, cramps, heaviness, pruritus and paresthesia) and six signs (edema, skin induration, hyperpigmentation, venous ectasia, redness and pain during calf compression). A summed total score of ≥ 5 indicates PTS, and > 14 or presence of ulcer indicates severe PTS.

Laboratory analysis

Blood was drawn at the time of diagnosis (before initiation of therapy), and on day 7, 30 and 90. Blood was collected in siliconized glass vacutainer tubes that contained 10 mL of 0.18% ethylenediaminetetraacetic acid (EDTA) for determination of IL-6 and CRP. IL-6 concentrations were determined with a commercially available enzyme-linked immunosorbent assay (Central Laboratory of the Netherlands Red Cross Blood Transfusion Service, Amsterdam, The Netherlands). These enzyme-linked immunosorbent assay (ELISA) kits are calibrated according to the standard of the National Institute for Biological Standards and Control (Potters Bar, UK). The detection limit of the assay is 3.0 pg mL−1. CRP was assayed with an immunoassay on a Vitros 250 (Ortho Clinical Diagnostics, Rochester, NY, USA), with a detection limit of 7 μg mL−1.

Statistical analysis

The relationships between inflammatory markers, non-invasive vascular measurements and PTS are reported as relative risk ratios (RR), with the associated 95% confidence intervals (CIs) in brackets, using the Poisson regression model [17–19]. This model allows direct estimation of relative risks adjusted for age, gender, body mass index (BMI) and other co-variates. Because of the high incidence of PTS, the use of this model is more appropriate than logistic models that only allow estimation of odds ratios.


Study population

A diagnosis of DVT was established in 184 patients, of which 68 were excluded for the following reasons: malignancy (n = 16), previous DVT (n = 6), immobilization (n = 5), recent surgery (n = 4), use of anti-inflammatory medication (n = 3), severe underlying medical condition (n = 3), recent trauma (n = 2), pregnancy (n = 2), bilateral thrombosis (n = 1), age below 18 (n = 1), already started with LMWH (n = 1) and a combination of exclusion factors (n = 24). Out of the 116 eligible patients three patients refused participation and three patients were lost to follow-up. Therefore, data of 110 patients were available for the final analysis. Among the participating patients, the risk factors for DVT were immobilization (n = 33), recent trauma (n = 12), oral contraceptive use (n = 23), surgery (between 4 weeks and 3 months prior) (n = 11) and positive family history (n = 36). The patient characteristics are shown in Table 1.

Table 1.   Baseline characteristics of the patients (= 110)
  1. DVT, deep vein thrombosis; BMI, body mass index.

Mean age at diagnosis (range)53.8 (25–86)
Male59 (54%)
Location of DVT
 Distal13 (11.8%)
 Proximal97 (88.2%)
 Left63 (57%)
 Right47 (43%)
BMI (SD)25.6 (4.9)
Duration of symptoms in days6.8 (1–90)

Il-6 and CRP

The median IL-6 plasma concentration was 7 pg mL−1 on the day of diagnosis, and below detection limit at day 7, 30 and 90. The 80th percentile of baseline IL-6 concentration was 15 pg mL−1. Among all patients the median plasma concentration of CRP was 20.5 mg L−1 on the day of diagnosis, 8 mg L−1 at day 7, and below detection limit after day 30 and 90. The 80th percentile of baseline CRP concentration was 54 mg L−1.

VOR, TS and reflux

VOR was elevated (> 0.8 mmHg × min/%) in 65 (61%) patients on day 7, in 51 (48%) patients on day 30 and in 33 (30%) patients on day 90. The median VOR declined in the first 3 months from 1.2 mmHg × min/% in the first few days after diagnosis to 0.8 and 0.6 mmHg × min/% after 1 and 3 months, respectively. Median TS scores declined from 5 (SD 3.4, range 0–13) on day 7 to 3 on day 30 (SD 2.7, range 0–10) and to 1 (SD 2.3, range 0–10) on day 90. High TS ≥ 3 were found in 65% of the patients on day 7, 50% on day 30 and 30% after 90 days. Reflux was present in 52% of the patients after 90 days. In 29% of the patients, reflux was present in one or two vein segments, in 15% in three or four vein segments and in 9% in five or more vein segments.

Incidence of the PTS

All patients were followed for over 1 year. Twelve patients experienced recurrent DVT, but only four recurrences occurred in the first year. After 1 year, the overall incidence of PTS was 36.7% according to the CEAP classification and 34.5% to the Villalta scale. Mean Villalta score was 4.1 (SD 3.8, range 0–19).

Compliance of stockings was good in both patients with and without PTS. The median numbers of days that stockings were worn during a week were 6.3 (SD 1.6) for patients with PTS and 6.4 (SD 1.6, P = 0.63) for patients without PTS.

Relative risks

On the day of diagnosis, increased levels of the inflammation markers IL-6 and CRP were related to higher outcomes of VOR after 3 months [IL-6: 2.4 (1.5–3.9) and CRP: 1.4 (1.1–3.3)]. IL-6 was also related to TS after 3 months [1.5 (1.1–2.1)]. For reflux, no relation with levels of IL-6 and CRP was found [IL-6 1.0 (0.6–1.6) and CRP 1.1 (0.7–1.8)]. All RR are shown in Table 2. Relative risks were not different when the group of patients with recurrent DVT was excluded.

Table 2.   Relation between inflammatory parameters and non-invasive venous examinations

Relation between two variables


95% CI

95% CI
  1. *Adjusted for age, gender and body mass index (BMI). Adjusted for age, gender, BMI and other soluble factors [interleukin-6 (IL-6), C-reactive protein (CRP), D-dimer]. VOR, venous outflow resistance; TS, thrombosis score.

IL-6 > p80VOR day 7 >–––1.7
IL-6 > p80VOR day 30 >–––2.2
IL-6 > p80VOR day 90 >–––3.9
IL-6 > p80TS day 71.21.1–––1.3
IL-6 > p80TS day 301.31.1–––1.5
IL-6 > p80TS day 901.51.1–––2.1
IL-6 > p80reflux day 901.00.6–––1.6
CRP > p80VOR day 7 >–––1.9
CRP > p80VOR day 30 >–––1.7
CRP > p80VOR day 90 >–––2.4
CRP > p80TS day 71.00.9–––1.2
CRP > p80TS day 301.00.8–––1.3
CRP > p80TS day 901.00.7–––1.4
CRP > p80reflux day 901.00.6–––1.8

After 1 year, seven patients had complete obstruction of at least one vein segment. In these patients, levels of IL-6 and CRP were higher at the time of diagnosis than IL-6 levels in patients without complete obstruction. On day 1, mean levels of IL-6 for patients with and without complete obstruction were 16.2 (SD 14.1) and 10.3 (SD 10.3), and mean levels of CRP were 51.1 (SD 37.0) and 29.4 (SD 28.5) for these sets of patients. For days 7, 30 and 90, there were no differences in levels of inflammation markers. Thirty-eight patients had complete recanalization and no reflux after 1 year. Levels of IL-6 and CRP showed no differences compared with patients with vein abnormalities.

Elevated outcomes of VOR, TS and reflux after 90 days were related to a higher incidence of PTS after 1 year. Relative risks after 90 days were 2.1 (1.2–3.7), 2.0 (1.1–3.7) and 1.7 (1.0–2.2) for VOR, TS and reflux, respectively, according to the CEAP classification and 2.1 (1.2–3.8), 2.1 (1.1–4.1) and 1.7 (0.9–3.0) using the Villalta scale (see Table 3).

Table 3.   Relation between non-invasive venous examinations and PTS after 1 year (CEAP-classification ≥ 3 and Villalta ≥ 5)

Relation between two variables

95% CI

95% CI

95% CI
  1. *Adjusted for age, sex and BMI. Adjusted for age, sex, BMI and other soluble factors (IL-6, CRP, D-dimer). VOR, venous outflow resistance; TS, thrombosis score; CEAP, clinical, etiologic, anatomic and pathophysiologic classification; Vi, Villalta classification for PTS.

VOR day 7 > 0.8PTS (CEAP ≥ 3)3.51.6–––5.1
VOR day 30 > 0.8PTS (CEAP ≥ 3)2.71.5–––4.0
VOR day 90 > 0.8PTS (CEAP ≥ 3)2.61.6–––3.7
TS day 7PTS (CEAP ≥ 3)5.30.8–––12.9
TS day 30PTS (CEAP ≥ 3)3.01.0–––5.4
TS day 90PTS (CEAP ≥ 3)2.21.1–––3.7
RefluxPTS (CEAP ≥ 3)1.91.1–––2.8
VOR day 7 > 0.8PTS (Vi ≥ 5)2.31.2–––7.5
VOR day 30 > 0.8PTS (Vi ≥ 5)1.60.9–––4.1
VOR day 90 > 0.8PTS (Vi ≥ 5)1.71.0–––3.8
TS day 7PTS (Vi ≥ 5)4.90.7–––41.1
TS day 30PTS (Vi ≥ 5)2.91.0–––10.1
TS day 90PTS (Vi ≥ 5)1.50.9–––4.1
RefluxPTS (Vi ≥ 5)1.71.0–––3.0

The association of IL-6 and CRP with PTS was weak using the CEAP classification with RR of 1.2 (0.7–2.2) and 1.8 (0.9–3.3) and absent using the Villalta scale 0.6 (0.2–1.4) and 1.2 (0.6–2.5), respectively (see Table 4).

Table 4.   Relation between inflammatory parameters and PTS after 1 year (CEAP-classification ≥ 3 and Villalta ≥ 5)

Relation between two variables

95% CI

95% CI

95% CI
  1. *Adjusted for age, sex and BMI. Adjusted for age, sex, BMI and other soluble factors (IL-6, CRP, D-dimer).

IL-6 >  p80PTS (CEAP ≥ 3)1.40.8–––2.2
CRP > p80PTS (CEAP ≥ 3)1.30.8–––3.3
IL-6 > p80PTS (Vi ≥ 5)0.60.2–––1.4
CRP > p80PTS (Vi ≥ 5)0.90.5–––2.5

INR values of the patients were below two in only 9% of the time. Mean INR value was 2.84 (SD 0.32). Adjustment for INR did not change the results.


To our knowledge, this is the first prospective study in DVT patients to determine the relation between inflammation in the acute phase of DVT and PTS as a long-term complication. We have demonstrated that in patients with DVT, elevated concentrations of IL-6 and CRP at baseline were associated with elevated VOR after 3 months and for IL-6 with high TS. No associations were found for reflux. Second, subjects with elevated VOR, TS and reflux 3 months after DVT had a higher risk of developing PTS after 1 year, which is in agreement with our previous study [5]. The association of IL-6 and CRP with PTS after 1 year was weak using the CEAP classification for PTS and absent according to the Villalta scale. Although the relation between inflammation and PTS is not yet fully elucidated, the results of our study suggest that inflammation might play a role in incomplete thrombus clearance, venous outflow obstruction and the future development of PTS.

Based on previous results and the current study, it is likely that two processes are involved in the development of PTS: (i) persistent obstruction (measured as VOR) because of incomplete thrombus clearance and venous wall fibrosis as a result of the inflammatory response in both the thrombus and the vein wall; and (ii) reflux as a result of direct damage to the valves by the thrombus itself or by entrapment in the fibrotic process of the vein wall. Both obstruction and reflux result in venous hypertension and PTS (see Fig. 1).

Figure 1.

 Hypothetic development of post-thrombotic syndrome (PTS): deep vein thrombosis (DVT) initiates an inflammatory response resulting in (incomplete) thrombus clearance [measured as thrombosis score (TS)], vein wall changes and fibrosis resulting in elevated venous outflow resistance (VOR). Direct mechanical damage to the valves might cause reflux of the valves. Both persistent obstruction and reflux because of valvular insufficiency lead to venous hypertension and PTS.

Interestingly, in our study, the detection of an association between inflammation and PTS as an endpoint depended on which clinical classification system was used. It is important to bear in mind that there is no gold standard or test to diagnose PTS. Although several classification systems have been developed and evaluated, their reliability, validity or responsiveness to change and their use for the routine clinical monitoring of DVT patients have not been fully assessed [20]. In a study in 124 patients with previous DVT, all PTS classifications showed a positive association with the reference standard ambulant venous pressure (AVP), but both the correlation of the CEAP classification and the Villalta scale (also referred as Prandoni scale) with AVP was moderate (0.22 and 0.27, respectively) and the correlation between the CEAP and Villalta scale was 0.42 (0.29–0.56) [21]. Furthermore, AVP showed considerable overlap in the different clinical groups. The CEAP classification has been widely implemented and is based on objective symptoms only and can serve as the basis for initial evaluation of PTS, but is relatively static for skin changes in time. The Villalta scale also includes subjective signs, and has no specific criteria for grading of the individual elements of the score. So, with the absence of a gold standard test, measuring outcomes of PTS is complex. The relation between inflammation and PTS can also be influenced by other variables determining PTS, such as intensity of anticoagulation [22] and the use of compression stockings, as the use of compression stockings prevents the development of PTS [4]. In this study, the intensity of anticoagulation was good, and compliance with wearing the stockings was high in all patients. Taking into account the above considerations, we conclude that there is a weak association between inflammation and PTS.

Moreover, the hypothesis that the inflammatory response leads to vein wall changes is supported by animal studies. Resolution of DVT resembles wound healing with early influx of neutrophils and monocytes, thrombus neovascularization, collagen deposition and late fibrosis leading to increased vein wall stiffness [23,24]. Inhibition of inflammation decreased the vein wall injury and stiffness independent of thrombus size [25]. Patients with DVT might benefit from treatment with anti-inflammatory drugs or longer treatment with LMWH, as LMWHs possess anti-inflammatory properties distinct from its anticoagulant properties [26], and LMWH improves vein wall re-endothelialization after DVT [27]. This hypothesis should be confirmed in future intervention studies.

A few issues about our study require comment. First, the number of patients excluded was high. The exclusion criteria were chosen to preclude the introduction of bias owing to detection of elevated levels of IL-6 and CRP as a result of other disorders than DVT and to ascertain that they represent the response to the thrombus. Second, 12 patients experienced recurrent DVT, the most important identified risk factor for PTS, but only four recurrences occurred in the first year. Exclusion of these four patients as the total of 12 patients with recurrent DVT, did not influence the results of the study.

In conclusion, although the relation between inflammation in DVT and PTS is not yet fully elucidated, the results of this study suggest that inflammation might play a role in incomplete thrombus clearance and vein wall changes resulting in PTS after 1 year. This suggests that a decrease in vein wall inflammatory response by anti-inflammatory medication or LMWH for a longer period of time may result in a decline in the manifestations of PTS.


The authors thank C. van der Vleuten (dermatologist) for assistance in the examinations of the patients after one year, and M. Albers-Akkers, Director of the Anticoagulation Outpatient Clinic Nijmegen. They also thank G. Pesman and J. van der Ven-Jongekrijg for determination of IL-6 and CRP. This study was funded by grant number 2001.133 from the Netherlands Heart Foundation.

Disclosure of Conflict of Interests

The authors state that they have no conflict of interest.