When defining PTS by four mandatory and predefined clinical criteria in line with clinical practice, we found that younger age and a higher baseline Villalta score were independent predictors of long-term PTS in patients with a prior high proximal DVT.
Our findings are likely to contribute to the limited knowledge of predictors for long-term PTS, and identifying patients at risk is an important step towards developing improved treatment strategies for the prevention of PTS (16, 45). Also, our findings may indicate that PTS remains a common chronic complication for up to ten years following a first-time high proximal DVT. Younger age was an independent predictor of PTS which is in contrast to previous studies reporting that older age is a predictor for PTS (1, 2, 44, 46, 47). Lower leg comorbidity is likely to be more common in older patients, and may have contributed to confounding in previous studies. This possible confounding was avoided in our study as obvious comorbid leg symptoms and signs excluded PTS according to the predefined criteria.
The baseline Villalta score is likely to indicate the symptom burden at the time of DVT diagnosis, and a higher baseline score was an independent predictor of long-term PTS. Others have also found that a higher baseline Villalta score predicts a less favorable long-term outcome. Rabinovich et al. developed a prediction model for PTS based on the SOX study and found that patients with a Villalta score ≥ 4 at baseline, in addition to pelvic DVT and body mass index ≥ 35 kg/m2, predicted increased risk of PTS (48). In addition, Kahn et al. found that higher Villalta score at one-month post-DVT was a strong predictor for PTS (2).
Previous studies have also found iliofemoral DVT (2, 44), recurrent ipsilateral DVT (2, 39, 46, 49), and high body mass index (BMI)(35) as strong predictors for PTS (44, 48). DVT with pelvic vein involvement were included in our multivariate model but was not a significant predictor in our study. The incidence of recurrent ipsilateral DVT was low, and data on baseline BMI was collected only in a subset of patients. Thus, neither of these two variables could be evaluated in the present study.
The introduction of CDT for PTS prevention was based upon the “open vein hypothesis”, where an accelerated thrombus removal is thought to prevent venous valve incompetence and incomplete recanalization (3, 50). At the six months follow-up of the CaVenT trial (42) the lack of iliofemoral patency and deep venous reflux were found to be independent predictors of PTS in the treatment group receiving CDT; and PTS was defined by the Villalta scale (40). Prandoni et al. reported an increased risk of PTS in patients with persistent venous obstruction or popliteal reflux at six months follow-up (24). In the current study, neither lack of iliofemoral patency nor deep venous reflux came out as independent predictors of PTS. This could be due to selection bias as only 52% of the still eligible patients were included, the lack of power, or the use of a different definition of PTS.
It is acknowledged that the clinical presentation of PTS is non-specific and conditions as primary venous insuffiency, trauma, central venous hypertension, and arthrosis may present with similar clinical manifestations (25, 31). Moreover, reported predictors of PTS are overlapping with predictors of primary venous insufficiency (29, 33). Galanaud et al. used data from the REVERSE study to assess risk factors for PTS in patients with a first unprovoked proximal DVT who were free of clinically significant primary venous insufficiency with an effort to remove biased evaluation of PTS. After excluding patients with primary venous insufficiency, only obesity remained an independent predictor of PTS (35). When assessing PTS by the four clinical criteria, we similarly tried to exclude patients with complaints likely to be explained by lower limb comorbidity. Previous studies have shown that Villalta scores in the ipsilateral and contralateral legs are strongly correlated, indicating that cases considered as PTS may reflect pre-existing chronic venous disease (16, 25, 26, 51).
A major limitation of this study includes the PTS assessment with no previously validated diagnostic tool. Hence, our findings can be considered as preliminary and hypothesis generating. Another limitation was that one study investigator performed all study visits and assessments. As only 52% of the eligible patients gave consent to participate in the current study there is a possibility for selection bias, however the participants and non-participants did not differ except for more leg comorbidity among non-participants (31). Comparisons between DVT with and without pelvic involvement in our material are uncertain, as pelvic involvement was not routinely assessed in patients randomized to the control group (3). Moreover, we were not able to assess BMI as patients’ height and weight were not systematically collected at baseline. A strength of the current study is the long follow-up time.