Introduction
Deep vein thrombosis (DVT) is a common condition. Its diagnosis is particularly important because of the associated morbidities and mortality, and because of the risk of pulmonary embolus (PE), a frequent and sometimes fatal short term complication. Longer term sequelae include recurrent venous thrombo-embolism (VTE) and post-thrombotic syndrome (PTS)(1). Despite the large healthcare burden caused by DVT/VTE, both in the inpatient and outpatient setting, relatively little is known about longer term DVT outcome, particularly as it relates to below knee as opposed to above knee DVT.
In the diagnosis of DVT, the use of contrast venography (CV) has been superseded by Doppler ultrasound, due to ultrasound’s combination of ease of access, non-invasive nature and lack of requirement for ionising radiation(2,3). Although contrast venography enjoyed ‘gold standard’ status in detection of both proximal and distal DVT(4,5), ultrasound offers sufficiently high sensitivity of around 97% for diagnosis of proximal vein DVT(6,7). However for detection of isolated distal vein DVT, ultrasound has a poorer sensitivity of only around 73%(6), resulting in some cases of isolated below knee DVT going undiagnosed.
Because of the reduced sensitivity of US compared with CV at detecting below knee DVT it is less reliable for assessment of the long term outcome with regards to recurrent VTE and PTS in this patient population. To circumvent this shortcoming we have retrospectively reviewed a cohort of CV diagnosed isolated DVT prior to the replacement of CV by ultrasound imaging for diagnosis of DVT.
This study aims to compare the long-term morbidity particularly with regards to PTS of patients with isolated below-knee/distal DVT versus those with above-knee/proximal DVT but also looking at incidence of recurrent DVT in this population. In addition, mortality was compared between the two groups and a control group without DVT. Risk factors present at the time of diagnosis of DVT were also examined for a potential correlation with the subsequent development of PTS.
Materials & Methods
Between October 1995 and March 2003, 1073 symptomatic patients underwent CV for investigation of suspected DVT at a single teaching hospital. These patients had clinical and imaging data recorded in a thromboembolism database, including the nature of the study performed, site investigated, and distribution of any detected thrombi. A range of risk factors present at the time of presentation were recorded, including previous instances ofDVT and VTE (including DVT affecting the contralateral side), recent immobility, obesity, recent surgery, recent trauma, history of thrombophilia, previous stroke, smoking status, heart failure, previous MI, history of malignancy, pregnancy, use of the oral contraceptive pill, use of hormone replacement therapy and family history of DVT/PE. Venographic imaging was performed following a standard protocol, which involved cannulating a foot vein, and applying tourniquets at the ankle and just above the knee to maintain contrast in the deep veins. Contrast injection by rapid hand injection under screening guidance ensured all calf veins were visualized, and the superficial femoral vein and pelvic veins were imaged immediately after release of the above knee tourniquet. Standard images acquired included two images of the calf vein, an oblique with the foot turned outwards, one long view of the upper leg veins and one of the pelvic veins. If any uncertainty persisted, further focussed views of a region in question were obtained by screening. All studies were reviewed by a consultant radiologist experienced in venography review. Patients without sufficient demographic data to facilitate follow up were excluded. Patients who had DVT excluded by CV were used as controls for the purposes of mortality follow up. Given the relatively small numbers, a control group for morbidity was not felt to be appropriate. All patients with above knee DVT, from the popliteal vein proximally and including iliac DVT, were assigned to the proximal/above knee group. Ethics approval was granted by the regional Ethics Committee.
Ten year mortality of all cases was obtained from death records held by the National Records Office, with cause of death categorised under PE/DVT, Cancer, Infection, Cardiopulmonary Event, Stroke or Other.
Morbidity was measured by recording the incidence of recurrent ipsilateral DVT events through case note review, evaluation of the clinical report regarding the presence of recurrent rather than residual.non-resolving thrombus, and whether symptoms of PTS were apparent, defined using the Villalta scale(8) via retrospective evaluation of patient case notes from subsequent clinic attendances and hospital admissions. A firm diagnosis of PTS was recorded as “definite PTS” if venous ulceration was present, a formal clinical diagnosis of PTS had been made, or if there was evidence from the written clinical notes of signs and/or symptoms consistent with PTS as per the Villalta scale diagnostic criteria. Those with “possible PTS” had only one or two signs or symptoms of PTS as per the Villalta scale. A minimum interval period of six months was used to determine presence of PTS, as residual symptoms of DVT may be difficult to differentiate from developing PTS(9).
Statistical analysis was performed using a Kaplan-Meier regression curve for mortality amongst those with proximal DVT, isolated distal DVT and the controls, and a log-rank test was employed to compare distribution of survival times across the three groups. Pearson’s chi-squared formula was used to evaluate for the presence of any statistical difference between the groups, and was evaluated at p<0.05 threshold.
Results
1073 patients were assessed for the presence of DVT by CV. Of the 399 with proven DVT, 52 were not suitable for follow up due to insufficient information being available, either due to a lack of demographic information or the incomplete recording of results from CV. A group of 355 controls were taken from those cases negative for DVT following CV, and were subsequently matched for age and sex with the study group.
Of the 347 patients with a DVT diagnosed by CV between 1996 and 2003, 236 patients had proximal DVT, whilst 111 had isolated below-knee DVT. Tables 1 and 2 demonstrate age and sex demographics and risk factor frequency respectively. Subjects were followed up for a period of up to 16.5 years, with a median follow up of 11.8 years. This allowed calculations of 10-year survival as well as complication rate, as described below. Patients diagnosed with DVT were treated using the hospital's standard protocol of low molecular weight heparin followed by warfarinisation.
Mortality
The 5-year and 10-year survival in both groups with proven DVT versus the control group is observed using a Kaplan-Meier cumulative survival chart (Figure 1). The 5-year survival of the below knee thrombus group was 64.5%, whilst the above knee group's survival was 73.6%, which is not a significant difference using the log rank test (p=0.08). The control group had a 5-year survival of 71.2%. The 10-year survival of the below knee group was 51.8%, whilst the above knee group survival was 60.4%, again not a significant difference (p=0.11). The control group had a 10-year survival rate of 58.5%.
Recurrent DVT
In the 10 year follow up, recurrent DVT was diagnosed in 69 (19.9%) patients with DVT, with higher recurrence rates in the above knee DVT group (55, 23.3%) compared to the isolated below knee DVT group (14, 12.6%). Recurrent DVT had a relative risk (RR) of 0.54 (95% CI 0.32-0.93) for below-knee versus above-knee DVT.
Recurrent PE
In the 10 year follow up, recurrent PE was diagnosed in 22 (6.3%) patients, with marginally higher recurrence rates in the above knee DVT group (15, 6.4% of patients with above knee DVT) than those with isolated below knee DVT (7, 6.3%). Relative risk of recurrent DVT after below knee DVT versus above knee DVT was 0.992 (95% CI 0.422-2.36).
Post Thrombotic Syndrome
The precise point of onset of PTS is difficult to deduce from the clinical notes, however overall 51 (14.7%) and 43 (12.4%) of the DVT patients were diagnosed with PTS or had possible PTS respectively (Table 4) in the ten years following presentation. In terms of post-thrombotic syndrome in the below knee DVT group, 11 (9.9%) had definite PTS, and 10 (9.0%) had possible PTS. In the above-knee patients, 40 (17.0%) had definite PTS, and 33 (14.0%) had possible PTS. Relative risk for definite PTS was 0.54 (95% CI 0.27-1.11) for below-knee DVT versus the above-knee group. Results were similar for probable PTS (RR 0.56, 95% CI 0.28-1.26) and for probable and definite PTS combined (RR 0.52, 95% CI 0.30-0.90).
Risk factors at initial presentation associated with long-term sequelae
In the patient groups with below knee DVT, above knee DVT and both groups combined who developed definite PTS, no correlation between any of the risk factors (as described in Materials and Methods) ascertained at initial presentation significantly correlated with subsequent development of PTS. In patients with prolonged immobility prior to clinical presentation with below knee DVT, there was a statistically significant increase in recurrent DVT (RR 10.27, 95% CI 1.392-75.852). This was replicated in the above knee group (RR 1.71, 95% CI 1.05-2.77). No other risk factors were statistically associated with development of recurrent DVT.
Discussion
The aim of this study is to examine the long term morbidity and mortality of patients by comparing those with isolated distal DVT to those with above knee proximal DVT. Proximal DVT is known to have a significant recurrence rate and elevated incidence of PTS(9–11). We aimed to establish to what extent if any distal DVT is associated with recurrent VTE and the development of PTS, a finding which would be of importance in determining treatment pathways.
Although compression ultrasonography has replaced contrast venography as the universally accepted investigation for the presence of DVT given its ease of access, safety and lack of invasiveness(9,10), CV is widely considered to be the gold standard(12), with a historical role in DVT diagnosis if initial ultrasound (US) was negative despite a high clinical suspicion(6,13). Given the relative operator dependence and likely reduction in accuracy of CV as the use of ultrasound predominates, Computed Tomography (CT) venography has emerged as an alternative diagnostically equivalent(14) tool for DVT diagnosis, although Magnetic Resonance (MR) venography offers another option without the need for ionising radiation.
Debate persists regarding long term outcome of patients with proximal versus isolated distal DVT, with minimal therapy and extended periods of treatment both advocated(15,16). Distal DVT has traditionally been regarded as less severe given the reduced rate of embolisation(17). A recent prospective study has demonstrated patients with distal DVT have a reduced incidence of VTE recurrence versus those with proximal DVT(18), although pulmonary embolism recurrence is similar. However, the risk of recurrent DVT in male patients and in those with malignancy may be of note(19). Current guidelines recommend twelve weeks of anticoagulation, however six may be sufficient (20).
Post thrombotic syndrome is estimated to occur in 43-47% of patients following DVT(21), although higher rates are suggested in those with iliofemoral DVT(22), not withstanding the higher risk of DVT recurrence involving this region. Retrospective study reports that 30%, 10% and 3% diagnosed with DVT subsequently suffer from mild, moderate and severe PTS respectively(23). It is recommended PTS is not diagnosed before three months have elapsed since initial presentation with an acute DVT, given the difficulty in differentiating between acute and chronic symptoms(1). Estimation of the prevalence and diagnosis of PTS is important given its association with long term morbidity, with self reported quality of life indicators low and correlating with other chronic diseases(24).
Therapeutic practice at our institution at the time of this study was to anti-coagulate all cases of proven DVT however we cannot comment as to whether the anticoagulation in any cases was sub-therapeutic. Even despite optimal anti-coagulation, between 20% and 40% of patients suffer PTS in the first two years following presentation(10,21). Data on the use of compression stockings was not gathered given the difficulty in retrospectively identifying their prescription and reliably ensuring therapeutic adherence.
Given 18.9% of those patients with distal DVT subsequently suffered either definite or possible PTS and the relative risk for definite PTS for distal DVT versus proximal DVT is 0.54, this group of patients cannot be ignored as regards to vigilance for future evidence of the development of PTS. For the reasons previously outlined, this is also likely to be an underestimate of the true incidence. In the performing institution, patients are followed up by a haematologist for eight weeks following a thromboembolic event. Therefore early signs of PTS may simply be mistaken as residual to the initial DVT. Diagnosis may fall into the realms of primary care, where a lack of awareness of this condition, its complicated means of confirmation and time constraints may result in recording a timely diagnosis extremely challenging.
Although not reaching significance, a small difference in mortality between the proximal and distal DVT groups is demonstrated, with the 10-year survival of the below-knee group 51.8%, as compared with 60.4% for the above knee group and 58.5% for the control group. Survival in the group with above knee DVT is comparable to other series(25). In our series, the prevalence of cancer is much higher in the below knee group, a finding at odds with other data. 48% of patients with below knee DVT group had malignancy listed as a cause of death, as opposed to 28% in the above knee group. Coupled with the higher percentage of male patients in the below knee thrombus group and a slightly higher median patient age, this goes some way to explaining the slightly higher mortality in the below knee patient group. Also of note is the higher survival after 1 year within the above knee patients as opposed to the below knee group. Though unexpected, this highlights the fact patients being investigated for thromboembolic disease are likely to suffer significant co-morbidities and undergoing investigation for suspected DVT is in itself an indicator of poor prognosis. The significant correlation between the development of PTS and the patient's smoking history has been described before, albeit only in a pregnant population(26), and highlights the need for smoking cessation advice in this group of patients. The association between development of PTS and a history of recurrent DVT is recognised in other studies(27).
A number of studies have reviewed the sensitivity of whole-leg Doppler US versus more widespread proximal compression techniques(28–30). These studies are consistent in finding no significant difference in adverse outcomes between the techniques, however one study did record an 11.3% rate of technical failure(30) due to obesity, active infection, active cancer and immobilisation, risk factors inherent to this population. In addition, follow up was limited to three months, thus the incidence of PTS is likely to be underestimated.
There are a number of limitations of this study. The retrospective nature means the incidence of PTS is likely to be under-estimated, given the reliance on review of patient notes and thus both accurate record of clinical symptoms and signs, and also on patient presentation. Thus those with mild symptoms are the patients most likely to have been excluded. Given the range of clinical signs and symptoms associated with PTS, the clinical diagnosis is often not made, particularly if the patient is not subject to longer term follow up after treatment for a DVT. The lack of a simple test for PTS and the reliance on a grading scale of constellation of signs and symptoms adds to this difficulty. Use of the deaths register for mortality data is another source of inaccuracy, however this partially offset by the use of an electronic patient record in our area meaning patient records are complete and easy to access. The numbers of patients included are relatively small, and thus drawing finite conclusions around risk factors would require further examination. A potential confounding factor is that some patients have suffered a DVT of either leg in the past. The retrospective nature of this study meant further information is difficult to obtain, however the relatively small numbers mean the effect of this group is likely to be small, and within this group a proportion will have had contralateral DVT unconnected to this presentation event and subsequent follow up. In patients with a recurrent DVT, it is difficult to exclude a recurrent above knee DVT as the cause of subsequent PTS, however it remains that after an initial below knee DVT they have suffered a further VTE event.
As expected, morbidity in the form of PTS is greater in those patients with DVT extending above the knee, however a smaller but significant proportion of patients with isolated below knee DVT also develop signs and symptoms of PTS. Ultrasonographic diagnosis of isolated distal DVT remains more technically challenging but this study supports the importance of making the diagnosis and instituting treatment of isolated below knee DVT in order to minimise recurrent DVT and to attempt to reduce the subsequent development of PTS. Further prospective study is required to evaluate this.