Venous blood flow, thromboembolism and below knee cast immobilisation for trauma

Hickey, Benjamin 2017. Venous blood flow, thromboembolism and below knee cast immobilisation for trauma. MD Thesis, Cardiff University. Item availability restricted.

Preview	PDF - Accepted Post-Print Version Download (1MB) | Preview
	PDF - Supplemental Material Restricted to Repository staff only Download (674kB)

Abstract

Venous thromboembolism (VTE) has a background incidence of between 0.7 and 2.69 per 1000 per year (L. N. Roberts et al., 2013). Risk factors are either permanent or transient. Permanent risk factors include thrombophilia (80 x increase risk if homozygous for factor V Leiden), cancer (58 x increase risk if metastatic cancer or diagnosis within last 3 months), increasing age (risk doubles for each decade over age 40 years), family or personal history of deep vein thrombosis (2-3 x increase risk) and increasing body mass index (2 x increase for BMI > 35 kg/m2 in comparison with BMI <20 kgm2) (Y.-H. Kim & Kim, 2007) (Blom, Doggen, Osanto, & Rosendaal, 2005) (Anderson & Spencer, 2003) (Decramer, Lowyck, & Demuynck, 2008) (Holst, Jensen, & Prescott, 2010). Transient risk factors include surgery (165 x risk in first 6 weeks after total hip or knee replacement, equating to 2% symptomatic VTE rate) (Sweetland et al., 2009). Foot and ankle procedures including ankle fracture fixation, hindfoot fusion and 1st metatarsal osteotomy are associated with 18x, 8x and 2x increase VTE risk respectively (Jameson et al., 2011). Other transient risk factors include postpartum state (21-84 x increase in first 6 weeks), use of oral contraceptive pill or hormone replacement therapy (at least 2 x risk) and lower limb cast immobilization (Jackson, Curtis, & Gaffield, 2011) (Grodstein et al., 1996). Within 90 days of lower limb cast treatment, asymptomatic DVT affects between 4 and 40% of patients, symptomatic DVT affects 1 in 250, symptomatic pulmonary embolism affects 1 in 500, with fatal pulmonary embolism affecting 1 in 15,000 (Jameson et al., 2014). It is apparent that patients will therefore have a differing risks depending on their permanent and transient risks. The types of VTE include asymptomatic events, for which the relevance is not fully understood (often used in studies as a surrogate for symptomatic events). Symptomatic below knee DVT (approximately 20% propagate to become above knee) (Philbrick & Becker, 1988). Symptomatic above knee DVT (affecting popliteal vein or more proximal), which are 4 times more likely to occur (Baglin et al., 2010). Pulmonary embolism can also occur. The clinical relevance of DVT is that 6% of patients will have severe post thrombotic syndrome (venous ulceration, swelling, itching) at 10 years after the event, with 66% of patients displaying some signs (Schulman et al., 2006). Uncomplicated DVT does not appear to impact on quality of life, however if DVT is complicated by post thrombotic syndrome, patients will have significantly reduced quality of life, mental and physical health. Simple non fatal PE reduces physical health and if it is complicated by pulmonary hypertension (affecting approximately 2%) it results in significantly reduced quality of life, mental and physical health (Ghanima, Wik, Tavoly, Enden, & Jelsness-Jørgensen, 2017) (Lubberts, Paulino Pereira, Kabrhel, Kuter, & DiGiovanni, 2016). In view that VTE has significant effects on patients quality of life, it is important to try and prevent it. In order to develop strategies for preventing DVT in patients with lower limb injury treated with leg cast, it is important to investigate the relative contributions of injury, stasis and immobility to thrombogenesis. I start by performing systematic review of the literature to determine whether thromboprophylaxis reduces symptomatic venous thromboembolism in patients with below knee cast treatment for foot and ankle trauma. A systematic review of randomised controlled trials of thromboprophylaxis in patients with foot and ankle injuries treated with cast immobilization was performed, searching MEDLINE and EMBASE from inception to June 2015 (B. A. Hickey, Watson, et al., 2016b). Outcomes of interest were VTE (asymptomatic and symptomatic DVT and PE) and bleeding. 3 reviewers used a data extraction form and assessed the literature according to the Cochrane risk of bias tool. Statistical analysis was performed using RevMan. 7 studies of chemical thromboprophylaxis were included, all except one used venography to assess for DVT, with one study using venous ultrasound. 2 studies reported on mechanical thromboprophylaxis, neither reported symptomatic DVT events. Neither study of mechanical thromboprophylaxis found a reduction in asymptomatic DVT in the intervention group. Funnel plot of studies of chemical thromboprophylaxis suggested no publication bias. Pooled symptomatic DVT occurred in 1.58% of patients in the control group, with 0.43% sustaining symptomatic PE. At meta analysis, symptomatic DVT was reduced in the low molecular weight heparin chemical thromboprophylaxis group (OR 0.29, CI 0.09-0.95). Chemical thromboprophylaxis did not influence PE. There was one non-fatal retroperitoneal haemorrhage (major bleed), which equated to 0.11% (1 in 886). Based on these findings, 11 symptomatic VTE events would be prevented for every 1 major bleed. These findings are comparable with the recent Cochrane review, which included 2 additional studies and a total of 2924 participants. Meta analysis found reported a reduction of VTE in the LMWH chemical thromboprophylaxis group (OR 0.40, 95% CI 0.21-0.76) (Zee, van Lieshout, van der Heide, Janssen, & Janzing, 2017). In order to develop strategies for prediction and prevention of VTE in patients with foot and ankle injury treated with cast immobilization, it is necessary to consider why venous thrombosis occurs in these patients. As previously discussed, patients may have permanent risk factors, which may influence hypercoagulability. The transient risk factors of injury and cast treatment may also influence risk by causing endothelial dysfunction and venous stasis (Virchow, 1856). Several important mechanisms for prevention of venous stasis have previously been found. Weight bearing is important; with Gardner et al (1990) reporting that 30ml of venous contrast was pumped out of the foot during weight bearing (Gardner & Fox, 1983). This is not always possible for a patient with foot and ankle injury treated with a cast, because they may be non-weight bearing. For patients who are non-weight bearing, it is still possible to influence venous flow. For example, Elsner et al (2007) previously found that movement of the 1st metatarsophalangeal joint increased popliteal vein flow from 13 to 39 cm/s (Elsner, Schiffer, Jubel, Koebke, & Andermahr, 2007). In patients without leg casts, intermittent pneumatic compression of the leg or thigh to prevent venous stasis was found to be effective in reducing DVT and PE in a meta analysis of over 16, 000 patients (RR 0.43, 95% CI 0.36-0.52) (Ho & Tan, 2013). It therefore seems that this is a viable mechanism. Furthermore, Whitelaw et al (2001) found that none of the IPC devices studied resulted in significantly better calf pump function when compared with simple passive or active ankle movements (Whitelaw et al., 2001). To assess the influence of toe and ankle movement on venous stasis, I examine the effect of these movements on venous velocities measured at the popliteal vein with ultrasound. To determine whether this is a viable strategy for prevention of DVT, I then assess the impact of application of below knee cast on venous velocities. In this proof of principle study, 20 healthy volunteers were recruited (B. A. Hickey, Morgan, Pugh, & Perera, 2014). All had measurement of calf pump function in the un-casted leg whilst seated, using ultrasound at the popliteal vein. Baseline and peak velocities were measured during active toe movement (dorsiflexion and plantarflexion) and during ankle movement (dorsiflexion and plantarflexion). A below knee cast was then applied and measurements were repeated. Mean resting baseline venous velocity was 10 cm/s, which remained unchanged when the below knee cast was applied. There was approximately 5-fold increase in venous velocities with active toe movement (mean 54 cm/s for toe dorsiflexion, mean 50 cm/s for toe plantarflexion), and 10 fold increase from baseline with ankle movements (mean 115 cm/s ankle dorsiflexion, mean 87 cm/s ankle plantarflexion). All were statistically significant. When the below knee cast was applied, there was no statistically significant decrease in the peak velocities achieved during movement excepting for ankle dorsiflexion (isometric), however this was still increased approximately 8 times compared with baseline (88 cm/s). It was therefore apparent that venous stasis did not occur when a below knee cast was applied to healthy volunteers and that active toe movement may have a role in preventing stasis in patients with injury, with subsequent reduction in DVT. To determine whether this is true I assess the effect of active toe movement on calf pump function and asymptomatic deep vein thrombosis in patients with foot and ankle injury treated with leg cast. In this prospective randomized controlled trial, patients between the ages of 18 and 60 years with acute foot and ankle injury treated with non-weight bearing cast were recruited (B. A. Hickey, Cleves, et al., 2016a). Patients were within 3 days of their injury and considered low risk for VTE after risk assessment. Those with additional risk factors were provided with LMWH and not recruited. Patients who consented, were randomized to either active toe movement (AToM) intervention group (advised to perform active toe dorsiflexion and plantar flexion 60 times every 6 hours minimum, but more often if possible). Patients were managed through the trauma clinic according to their injury then had assessment of calf pump function on removal of cast and assessment for lower limb DVT of both lower limbs using venous ultrasound. Interim analysis was performed after the first 100 patients were recruited. 78 patients completed the study, mean age was 37 years, 65% were male. 59% had leg cast for ankle fracture. Analysis of calf pump function revealed no significant difference between the intervention and control groups for any of the parameters, with mean baseline popliteal velocity of 7 cm/s in both groups, popliteal venous velocity during active toe dorsiflexion 44 cm/s (AToM) v 34 cm/s (control), p=0.36 and popliteal venous velocity during active toe plantar flexion 39 cm/s (AToM) v 32 cm/s (control) p=0.35). 27% of patients were found to have asymptomatic DVT, with no significant difference between groups. The important finding was that all asymptomatic DVTs occurred in the lower limb that had been injured and treated with cast. This basic finding had not been previously reported in the literature. It was important to determine whether this finding could be attributable to venous stasis alone, or whether general immobility or the injury itself had any significant role. To answer this, I assess the association between patient mobility and development of asymptomatic DVT. As part of the AToM study, a triaxial accelerometer (MOVBand) was attached to the leg cast at time of recruitment to the study. This was removed in the trauma clinic at the first appointment and the first 5 days of accelerometer data was extracted. Unpaired t test was used to determine statistical significance between group means for patients who did and did not develop asymptomatic DVT. 78 patients completed the AToM study, 10 patients were excluded from accelerometer data analysis (4 trackers lost, 6 failed to record any data). There was no significant difference in accelerometer data between patients who did and did not sustain asymptomatic DVT. Average moves were 1057/day (no DVT group) vs 1005/day (DVT group), p=0.85. Average steps were 877/day (no DVT group) vs 825/day (DVT group), p=0.82. In view of this, it appeared that mobility of patients during the first week of cast treatment for injury did not predict subsequent finding of asymptomatic DVT. In view that all DVT’s occurred in the lower limb that had been injured and treated in cast, it was apparent that local factors such as venous stasis or tissue injury were more important than general patient mobility, otherwise it would be anticipated that some DVT’s would have occurred in the uninjured, un-casted limb. To investigate the role of the injury I examine the association between biomarkers of coagulation and tissue injury with the outcome of asymptomatic DVT. As part of the AToM study, 3.5ml venous blood was taken at time of recruitment. Centrifuged plasma was stored at -70 degrees centigrade. After the last patient exited the study, plasma was analysed for levels of tissue factor, interleukin 6, vascular cell adhesion molecule 1 (VCAM-1) and D-dimer (B. A. Hickey et al., 2017). 77 patients were included, 1 patient did not provide blood sample. Analysis of results found no difference between levels of tissue factor, IL-6, VCAM-1 and D-dimer in groups who did and did not sustain asymptomatic DVT. Mean Tissue factor 23.9 pg/mL (no DVT group) vs 20.3 (DVT group), p=0.422. Median IL-6 3.9 pg/mL (no DVT group) vs 4.6 (DVT group), p=0.76. Median VCAM-1 553 ng/mL (no DVT group) v 496.8 ng/mL (DVT group), p=0.11. Median D-dimer 203.5 (no DVT group) v 236.0 (DVT group), p=0.49. I therefore appeared that severity of injury, measured using plasma levels of FT, IL-6, VCAM-1 and D-dimer could not predict which patients would develop asymptomatic DVT. This suggested that local factors of venous stasis might play a greater role in thrombogenesis than tissue injury. In summary, it appears that chemical thromboprophylaxis reduces the risk of symptomatic DVT in patients with foot and ankle injury treated with leg casts. Active toe movements can prevent venous stasis in healthy volunteers, however this does not appear to influence calf pump function or occurrence of DVT in patients with injury and leg cast. The important finding from this work is that asymptomatic DVT only occurs in the lower limb that has been injured and treated with leg cast. Biomarkers of coagulation do no appear to predict DVT in this patient group. Similarly, general patient mobility measured objectively with accelerometer during the first week of cast treatment does not appear to be associated with development of DVT in patient with foot and ankle injury treated with leg casts.

Item Type:	Thesis (MD)
Date Type:	Completion
Status:	Unpublished
Schools:	Engineering
Uncontrolled Keywords:	Venous Thrombosis; Thromboprophylaxis; Deep Vein Thrombosis; Tissue Factor; Biomarkers; Cast.
Date of First Compliant Deposit:	8 January 2018
Last Modified:	18 May 2021 14:28
URI:	https://orca.cardiff.ac.uk/id/eprint/107891

Actions (repository staff only)

Edit Item