Early 2000 MRI dti [11]
Numbers
30–76
60–305
75–103
102
2,518–2,860
Any DVT (proximal)a
53–75 %
50–73 %
22–34 % (16)
40 % (18)
17–21 % (10–12)
Clinical DVT
32–36 %
7 %
3 %
4–6 %
Any PEa
11 %
Clinical PE or autopsy
13–16 %
7–20 %
0–9 %
5 %
1–2 %
From 2000 onward, studies have tended to be larger multicentre trials on a less selected group of patients, usually using compression ultrasound. In patients not on VTE prophylaxis in the control arm of these interventional trials, the incidence of detected DVT was now lower at around 10–20 % (proximal 10 %, all DVT including distal 20 %), with clinical incidence being 1–3 %. In these studies, subclinical PE were not looked for, and the clinical incidence of PE was around 1–3 % [32, 33].
It may well be that greater emphasis on acute care of stroke, which includes better hydration and early mobilisation, has reduced the incidence of VTE, but it could also be said that compression ultrasound may not pick up all the DVT using older methods like radioiodine-labelled fibrinogen or venography. A novel study using magnetic resonance direct thrombus imaging (MRDTI) in 102 patients from 1999 to 2000 found the incidence of subclinical proximal DVT to be 18 %, and all DVTs (proximal and distal) to be 40 %, with a clinical incidence of 3 %. The study had the advantage of also picking up subclinical PE, and this was found in 12 % of patients with a clinical incidence of 5 % [11]. It is interesting that the clinical incidence of PE was higher than that of DVT, which is often not the case. The pick-up rate with MRDTI for DVT would be almost as good as venography and closer to the true incidence, and the same may be said for PE.
Nevertheless, it still could be argued that improvements in acute management of stroke even since the turn of the century have brought the incidence of VTE down. Note the clinical incidence of PE from the MRI study at 5 % is higher than the 1 % found in the latest studies.
Intracerebral Haemorrhage
As mentioned in the Introduction, studies suggest that VTE may occur more frequently in intracerebral haemorrhage (ICH). This may be a result of more severe weakness, depressed level of consciousness, and the avoidance of routine antiplatelets in the ICH. Retrospective studies comparing the clinical incidence of VTE in ischemic stroke (IS) and ICH have consistently found a higher rate of DVT and PE in ICH than in IS. In one cohort, the rate of DVT was 1.9 % of 1,126 patients with ICH vs. 0.5 % of 15,599 patients with IS, the rate of PE 0.4 % vs. 0.1 % [34]. In another, larger, cohort DVT occurred in 1.37 % of 1,606,000 ICH vs. 0.74 % of 14,109,000 IS, PE 0.68 % vs. 0.51 % [13]. In patients with hemorrhagic stroke initially included in the International Stroke Trial, it was found that patients with ICH had a higher rate of clinically diagnosed PE than patients with IS (1.3 % vs. 0.7 % P = 0.06) [35].
Below-Knee, Distal, or Calf DVT
The incidence of below-knee, distal, or calf DVT from control groups without any planned prophylaxis in recent trials has been between 7 and 9 % [32, 33]. And these constitute 40 % of all DVT, though it is likely that some are missed by compression ultrasound. In the MRDTI study, the incidence of below-knee DVT was 22 % and constituted 55 % of all DVT [11]. The figures above are the total incidence; the symptomatic or clinical incidence of below-knee DVT is around 1–3 %, and probably just slightly less than symptomatic proximal DVT. Without treatment, proximal extension into the popliteal vein occurs in 20 % of calf DVT after 2 weeks [11]. There is a risk of PE resulting from calf DVT either because of propagation or embolism from the calf itself. There are no studies of the risk of PE in stroke patients, but in other patients with calf DVT, the risk of symptomatic PE is up to 1 % and asymptomatic PE up to 6 % [36].
Pelvic and Inferior Vena Cava DVT
DVT in the pelvis and inferior vena cava (IVC) after stroke being less common than DVT in the legs have been less studied. Also, the usual imaging methods used to pick up DVT would not have easily detected pelvic and IVC DVT. The overall incidence of these DVT in the stroke population is probably low. In the study using MRDTI to detect VTE, isolated pelvic DVT was found in 2 % of patients and constituted 5 % of all VTE [11]. In selected groups of stroke patients, the incidence can be higher; for instance, in patients with cryptogenic stroke, the incidence of pelvic DVT can be as high as 20 % [37]. In another selected patient group, autopsy series from medical patients with pulmonary embolism, the source of emboli was identified as the pelvic veins in 11 % and the IVC in 5 % [38]. Whilst these DVT can be an extension of femoral DVT, they can often occur alone. In a study using MRDTI in 44 medical patients with PE, the source of emboli was found to be the pelvis, and IVC in 4 (9 %) of the patients, with the thrombus being an extension of a femoral DVT in only 1 case (2 %) [39].
Upper-Extremity DVT
There a no studies looking at the incidence of upper-extremity (axillary, subclavian, and internal jugular) DVT in stroke. From experience, it does occur in the paralysed upper limb, and there are a few case reports [40], but it less common. It is surprising that DVT in the equally paralysed arm does occur less frequently than in the leg, and there is no adequate explanation for this. Some have pointed to the difference in venous anatomy of the arm or greater fibrinolytic activity in the arm [1, 41]. In the literature from unselected patients, it could account for around 4 % of DVT [40]. Whilst there are no studies in stroke, upper-extremity DVT carry a risk of PE with figures of 9–36 % in other groups of patients [40, 42].
Impact of VTE in Stroke (Clinical and Subclinical)
There is a high morbidity and mortality in patients with clinical VTE after stroke. It is the third (or fourth) leading cause of post-stroke mortality after the stroke itself, and secondary infection (and possibly cardiac disease) [43, 44]. The consequences of PE are not just death but significant morbidity, with debilitating symptoms of pain and shortness of breath, and further complications including pulmonary infarction and infection, pleural effusion and empyema, and in the longer term, pulmonary hypertension [45]. There is association with greater disability and longer length of stay [46]. In terms of the impact of clinical DVT, there is, of course, the risk of PE and death, but furthermore, there is the morbidity around limb pain, immobility, and later, the post-thrombotic syndrome [45, 47]. The prevention of VTE may not only improve mortality and morbidity but also recovery and rehabilitation, as it has been shown that DVT prophylaxis is one of the processes of care associated with good outcome, including reduced rates on institutionalisation [48].
In the past when the incidence of DVT after acute stroke was high, between 50 and 75 %, and anticoagulants were felt to be unsafe, the impact of VTE was very clear. There was a high clinical incidence, around half DVT were clinical and 22–30 % of patients with clinical and subclinical DVT went on to suffer a PE—a mixture of half clinical and half unsuspected post mortem diagnosis (13–16 % of the study population) and 25–50 % of those patients with PE died (3–8 % of the population). Of the patients with clinical PE, up to 30 % of patients died, but the majority of the patients with PE who died were found in autopsy and were subclinical [5, 6, 25, 26]. Understanding whether subclinical PE was the primary cause of death in these cases can be sought through post-mortem evidence.
Examining studies with post-mortem evidence, one reported 50 % of PE after stroke presented as sudden death, with confirmatory findings of massive PE including saddle emboli, which were likely to be fatal in all the cases [49]. The peak incidence for mortality from PE is felt to be 2–4 weeks after stroke onset, but can occur as early as the first 3 days [49, 50]. Looking at post-mortem series from the 1970s to 1980s of patients who die after acute stroke, the PE were detected in up to 13 %, but increased to 30 % for deaths in the second to fourth week, mostly undiagnosed before death [44, 51, 52], lending evidence that subclinical PE may be leading to death. The presence of PE could be underestimated even in such studies because autopsies are not carried out in all post-stroke deaths. In one heparin prophylaxis study from the 1980s on a selected group of older stroke patients, autopsies were carried out in over 80 % and the presence of PE was detected in 56 %. There was also a link with ante-mortem diagnosis of DVT, with PE being found in 76 % of deaths compared to 20 % of those without DVT. Most of the DVT were subclinical, and further evidence of relationship between subclinical DVT and PE-related deaths was that the group receiving prophylaxis had fewer DVT, 22.2 % vs. 72.7 %, fewer deaths, 22 % vs. 33 %, and in those who died fewer PE were found, 29 % vs. 70 % [28].
So in summary it does seem there is reasonable evidence that subclinical DVT was leading to subclinical PE, which in turn was contributing to post-stroke deaths. In other patient groups, it has been suggested that one-third to two-thirds of PE found at autopsy caused or contributed to the death [53, 54]. So at that time both clinical and subclinical VTE did seem to matter, and whilst subclinical PE may not have been the final cause of death in all cases, its presence may have been contributory, and in patients who do not die, the presence of PE may well be impeding recovery and rehabilitation [25, 48].
The incidence of VTE and death from that era was quite high, as prophylaxis was not a standard practice, and conventional treatment with anticoagulation—even for established VTE—was felt to be hazardous after acute stroke. There has been a fall in clinical DVT and PE after stroke in the last few decades, with greater use of antithrombotic medication and better acute stroke care [32, 33, 55, 56]. With clinical PE being about 1 %, the impact of VTE has been less evident, and whilst the significance of clinical PE is still high, with mortality rate of up to 40–50 % [57], the case for prevention is less overwhelming, especially if it involves some risk like prophylactic anticoagulation.
So prevention in this era is based around the benefits of prevention of subclinical VTE. Extrapolating from hospitalised medical patients, having subclinical events still matter, subclinical proximal DVT was associated with higher mortality than subclinical distal DVT, which again was higher than no DVT [58]. The MRDTI study mentioned previously gives a more recent perspective of subclinical post-stroke VTE, having the advantage of detecting not only subclinical DVT, but PE as well. In this unique study, DVT was found in 40 % of patients, proximal in 18 % of patients, and only 3 % of these DVT were clinical; PE was found in 12 %, 5 % clinical and 7 % subclinical, and attributed mortality from PE was 2 % [11]. The incidence and mortality from PE could have been higher, as patients with proximal DVT were anticoagulated when found, including those who were subclinical. Of the 5 % clinical PE, only 3 % had been recognised, and in the 3 % clinical DVT, only 1 % recognised.
It is felt that fatal PE usually arise from proximal DVT, and from studies in other groups of patients, especially postoperative patients, the mortality from PE in untreated clinical proximal DVT can be as high as 40 %. The mortality from PE in untreated subclinical proximal DVT is felt to be lower, at 5–15 % [47]. Using this MRDTI study as a model, the mortality of 2 % out of a total PE incidence of 12 % approximates to 20 % of patients with PE. Also, 2 % mortality out of a total proximal DVT incidence of 18 % equates to a mortality rate of just over 10 % in patients with proximal DVT. To simplify, it could be suggested that in patients with proximal DVT after acute stroke (which is mainly subclinical), PE occurs in 50 % or more, and 10 % or more may die. Even if the mainly subclinical incidence of proximal DVT may have fallen to 10 % in this era of hyperacute stroke management [32, 33], this still equates to PE rate of 5 % (the majority may be subclinical as the rate of clinical PE in recent trials is 1–2 %) and mortality of 1 % (some of which may be unrecognised as deaths due to PE). It is figures such as these which justify the prevention of VTE, otherwise many remain undiagnosed and untreated, with significant consequences, hidden or otherwise.
Prevention of VTE
Once the incidence of VTE in stroke patients was investigated for in the early 1970s and found to be excessively in high in immobile patients with attendant mortality, various forms of prevention have been tried to reduce the incidence of VTE. In addition, 5 % of stroke patients will have a previous history of VTE, putting them at greater risk regardless of whether there is paralysis and immobility [33]. There has been a debate over the best form of prophylaxis, with a different stance being taken on either side of the Atlantic, especially on using anticoagulants as thromboprophylaxis. The components of current clinical guidelines are reviewed below (Table 6.2) [59–73].
Table 6.2
Summary from the clinical guidelines of their guidance on prevention of VTE
UK | US | Europe | Australia | Japan | |||
|---|---|---|---|---|---|---|---|
RCP 2012 [59] | NICE 2009 [62] | ANSF 2010 [71] | |||||
Ischaemic stroke | Ischaemic stroke | Ischaemic stroke | Ischaemic stroke | Ischaemic stroke | Ischaemic stroke | Ischaemic stroke | Ischaemic stroke |
Early mobilisation and hydration mentioned (GCP) | Early mobilisationand no evidence to support or refute | Early mobilisation to prevent subacute complication (Class I, Level C) | Early mobilisation and hydration (Class IV GCP) | Early mobilisation and hydration (GPP) | |||
GCS should not be used as prophylaxis (Class 1, Level A) | IPC should be considered. (Grade A). Above knee GCS not recommended (Grade A) | Do not offer GCS | IPC for those who cannot receive anticoagulants (Class IIa, Level B then) | IPC first line or prophylactic UFH or LMWH (Grade 2B) GCS suggested against (Grade 2B) | GCS (Class IV GCP then) | Thigh-length GCS not recommended (Grade B) | Insufficient evidence for IPC and GCS at the time (Grade C1) |
Aspirin is recommended in the first 2 weeks (Grade A) | Aspirin reasonable in patients who cannot receive anticoagulants (Class IIa, Level A) | Antiplatelets should be used to prevent DVT/PE (Grade A) | Aspirin is not recommended for the prevention of PE in patients with ischaemic stroke | ||||
Prophylaxis with anticoagulants should not be used routinely after stroke. Where anticoagulation is needed for prevention LMWH is preferred to UFH | Anticoagulation in the first 2 weeks can cause haemorrhage and has no net benefit. Patient at particularly high risk can be given prophylactic heparin. LMWH recommended over UFH (in addition to IPC). After 2 weeks risk should be re-assessed | Prophylactic LMWH should be considered for those at high risk of VTE or UFH in patients with renal failure | Subcutaneous anticoagulants recommended in immobilised patients. Ideal timing of initiation not know (Class I, Level A) | Prophylactic UFH or LMWH but low dose LMWH suggested over UFH (Grade 2B) | Low dose UFH or LMWH should be considered for those at high risk of VTE (Class 1 Level A) | LMWH can be used with caution on selected patients at high risk or UFH if contraindicated (Grade B) | LMWH or UFH is recommended for patients with ischaemic stroke with paralysis of lower extremities but not routinely because of the risk of bleeding (Grade C1) |
Intracerebral haemorrhage | Intracerebral haemorrhage | Intracerebral haemorrhage | Intracerebral haemorrhage | Intracerebral haemorrhage | Intracerebral haemorrhage | Intracerebral haemorrhage | Intracerebral haemorrhage |
No specific recommendations | No specific recommendations | Initial use of IPC (Class I, Level B) treat hypertension | Initial use of IPC (Grade 2C then) GCS suggested against (Grade 2B) | GCS not recommended Initial use of IPC as first line—moderate | IPC or GCS or combination of the 2 (Grade B) | ||
Bleeding risk mentioned as contraindication to LMWH or UFH in ischaemic stroke | Prophylactic dose LMWH or UFH may be considered after cessation of bleeding 1–4 days from onset (Class IIb, Level B) | Prophylactic dose LMWH or UFH started between 2 and 4 days (Grade 2C). LMWH suggested over UFH (Grade 2B) | Insufficient evidence to make recommendation about anticoagulants—Low | Antithrombotics not recommended (GPP) | Low dose heparin can be considered for ICH patients without rebleeding 3–4 days after onset | ||
Early Mobilisation
Early mobilisation is almost universally mentioned in clinical guidelines as one of the first measures for the prevention of DVT (as shown in Table 6.2). There is, however, no direct evidence so far that it prevents DVT and it was a component in the immobility related adverse events that was not lower in the very early mobilisation group in the latest results of the AVERT trial [74]. In the time before stroke units were commonplace, DVT rates were historically lower in stroke units where early mobilisation was one of the components of effective stroke care. Prior to the recent cautionary results on very early mobilisation from the AVERT trial, early mobilisation was linked with other beneficial outcomes like increased independence, fewer pressure sores, fewer cases of pneumonia, increased psychological well-being, and a reduced length of stay [75–77] There were some concerns in the past that early mobilisation should stop after the diagnosis of DVT because of the risk of PE [78], but that was in the era when intravenous heparin and warfarin were used. In a recent meta-analysis of trials using LMWH, this was found not to be the case, and recommended once treatment started, early mobilisation should positively be encouraged, as there was a trend toward lower mortality and less progression of DVT [79].
Hydration
Dehydration is linked with DVT after acute stroke. A urea level above 7.5 mmol/l, a urea/creatinine ratio over 80, and serum osmolality of greater than 297 mOsm/kg were associated with greater odds of developing a DVT [8]. Therefore, it has been surmised that better hydration will prevent DVT, but there has not been a trial to prove it. In a Cochrane review of heamodilution trials mostly using dextran to unselected patients, there was a tendency to fewer VTE [80]. Changing osmolality in patients can be difficult, even in patients on intravenous fluids [81], but perhaps with closer monitoring of hydration parameters than is done routinely, this can be achieved [82]. Some of the current National Guidelines do recommend hydration as a first-line measure in stroke prevention [59, 69, 71].
Heparin, Heparinoids, and Low-Molecular-Weight Heparin in Ischaemic Stroke
Studies looking at the use of chemical thromboprophylaxis with heparin in stroke started as early as 1977 [26] and was in routine use as recommended prophylaxis after ischaemic stroke in United States since the 1980s [83]. In the UK, heparin prophylaxis was used in trials and selected patients, but the concern about secondary cerebral haemorrhage limited its use. In a meta-analysis from 1993, heparin was seen to show a significant 81 % reduction of DVT after acute stroke; there was also 58 % reduction of PE and 18 % reduction in mortality, but these latter two results were not significant. There was also a non-significant 12 % increase in haemorrhagic transformation of infarct [84]. The reasons these results were far from conclusive, apart from inadequate numbers, were that not all these trials were designed primarily to look at VTE thromboprophylaxis; a mixture of high and low doses of heparin were used, and the effects of secondary haemorrhagic transformation not fully understood. The large International Stroke Trial [85] had the numbers, but again was designed to look at antithrombotics as treatment for ischaemic stroke rather than prophylaxis against VTE. It had a factorial design and compared aspirin, together and against two doses of heparin (high-dose 12,500 units bd and low-dose 5,000 units bd) and placebo, started within 48 h of stroke onset, but did not have DVT as an endpoint. Clinical PE was an endpoint, and with heparin (combined high and low dose), there was a significant reduction from 0.8 to 0.5 % (p = 0.02) – 3 fewer PE/1,000 patients. There was, however, a significant increase in intracerebral haemorrhage (1.2 % vs. 0.4 % compared to not using heparin)—8 more cerebral bleeds per 1,000, which more than outweighed the benefit of reduction in PE. Aspirin did not show a significant increase in intracerebral haemorrhage (0.9 % vs. 0.8 %) and there was a slight trend to PE reduction 0.8–0.6 %, but this was not significant (p = 0.08). There was also no long-term survival or functional outcome benefit with heparin. When the outcomes with the two doses were subdivided in subgroup analysis, the protection against PE was associated with higher dose 12,500 units bd (0.5 % vs. 0.9 % control), but so was the haemorrhagic risk (1.8 % vs. 0.3 % control); the lower dose 5,000 units bd had the similar PE incidence as aspirin (0.8 % vs. 0.7 %) and intracranial haemorrhage (0.7 % vs. 0.5 %). Both doses demonstrated similar protection against ischaemic stroke as aspirin. The combination of lower dose of heparin 5,000 units bd and aspirin, which is sometimes used as VTE prophylaxis today, did seem to prevent early recurrent ischaemic stroke (2.1 % vs. 3.5 %) without a significant excess in intracerebral haemorrhage over aspirin alone (0.8 % vs. 0.5 %), but did not show a significant reduction in PE, 0.5 % vs. 0.7 %. Therefore, from these findings, the use of anticoagulation early after acute ischaemic stroke was discouraged in the UK guidelines.
In the US before and after IST, unfractionated heparin (UFH) and then low-molecular-weight heparin (LMWH) continued to be used immediately after acute ischemic stroke as thromboprophylaxis against VTE, and subsequent systematic reviews [86–88] continued to suggest that anticoagulants did prevent VTE, but with the risk of intracerebral haemorrhage. When the reviews were done combining high and low doses of UFH and LMWH, the conclusion was often that the benefits of VTE prevention were outweighed by the risk of symptomatic intracranial haemorrhage (sICH). In one systematic review, the low doses and high doses were separated, it was shown that a lot of the bleeding risk was associated with the high doses of heparin or LMWH, and that the lower doses had less of a bleeding risk but also less thromboprophylactic benefit, with low doses of UFH being shown to prevent DVT but possibly not PE. The best risk profile seemed to be with LMWH, with a significant reduction in DVT OR 0.34 (0.19–0.59) and PE OR 0.36 (0.15–0.87) and non-significant rise in sICH OR 1.39 (0.53–3.67) [88].
In the PREVAIL trial, LMWH was seen to be superior to UFH; Enoxaparin reduced the risk of venous thromboembolism by 43 % compared with UFH (10 % vs. 18 %; relative risk 0.57, 95 % CI 0.44–0.76, p = 0.0001), with no difference in symptomatic intracranial haemorrhage, 1 % vs. 1 % [89]. A criticism was that this trial was comparing two heparins, probably combined with aspirin which was in common use as stroke prophylaxis by then, and had no placebo arm. Since IST, a trial has not been done to look specifically at aspirin and low-dose anticoagulant versus aspirin and placebo, which is what is practised today as VTE prophylaxis in some parts of the world. In different countries, either it is taken as a given that anticoagulation is the standard prophylaxis, or that is generally discouraged unless under special circumstances and other methods are being looked at. Another point made about the PREVAIL trial was that there was no difference in survival, despite there being not only a difference in asymptomatic DVT, but there was also a difference in total numbers of symptomatic DVT and PE, 1 vs. 4 DVT and 1 vs. 6 PE, in favour of Enoxaparin (the numbers being small, the differences in symptomatic VTE did not reach significance). There are still concerns that only the symptomatic VTE events matter, and the symptomatic PEs prevented did not outweigh symptomatic extracranial haemorrhage in this trial and did not outweigh an increase in sICH when compared to placebo in other trials [90].
If prophylactic anticoagulation is used it should be continued for at least a month in patients with an immobile limb, for although the majority of DVT develop early, a significant proportion can occur up to a month (26–30 days) after the event, and extended prophylaxis beyond 10–14 days has been shown to reduce this [10, 91]. In stroke patients from the EXCLAIM Study, extended prophylaxis reduced VTE from 8 to 2.4 %, with one fatal PE in the placebo group, this albeit with an increase in major bleeding (1.5 % vs. 0 %), and one fatal intracranial bleed in the treatment group. Prophylactic anticoagulation may not need to be continued long term and it may not be practical to do so. There is evidence from chronic paraplegics that after some time has elapsed from the acute event, the risk of DVT is a lot less, possibly due changes in fibrinolytic activity in the limbs [41].
Heparin, Heparinoids, and Low-Molecular-Weight Heparin in Primary Intracerebral Haemorrhage
With the invention of the CT scan in 1972 and the first clinical head scans only being installed in the mid-1970s, the early heparin trials in the 1970s did not routinely use CT head scans, and thromboprophylaxis was given to strokes in general. With the more widespread use of CT in the 1980s, distinguishing between ischaemic stroke and intracerebral haemorrhage (ICH), anticoagulant prophylaxis was not used early in ICH, and in some countries not at any time, and most of the trials of anticoagulant prophylaxis in acute stroke have concentrated on ischaemic strokes. The US guidelines did allow its use after 48–96 h, based on evidence that it did prevent PE without increase in the number of patients with rebleeding [65, 68, 92], and there has been a recent study from a stroke registry showing that the use of prophylactic anticoagulation within 7 days did not seem to cause extension of the intracrebral haemorrhage in a majority of patients [93]. A meta-analysis from 2011 again showed a significant reduction in PE with anticoagulant prophylaxis, but not DVT, with a non-significant trend toward increased survival, but also a trend toward greater hematoma growth [94]. There have really not been any large well-designed controlled trials of anticoagulation prophylaxis in ICH. Other guidelines, including those of the UK, do not recommend its use at any time in intracerebral haemorrhage.
The Novel Oral Anticoagulation Agents as Prophylaxis in Acute Stroke
Over the last few years, several new oral anticoagulation agents have been introduced which have been in situations traditionally occupied by LMWH and warfarin, such as the prevention and treatment of VTE and the prevention of stroke in atrial fibrillation. The novel oral anticoagulants (NOACs) can be separated into direct thrombin inhibitors, such as Dabigatran, and direct Factor Xa inhibitors, such as Rivaroxaban and Apixaban. In the sphere of VTE thromboprophylaxis, these NOACs have been shown to be superior to LMWH at preventing VTE in post-operative orthopaedic patients without an increase risk of bleeding [95]. As yet, there are no trials of VTE prophylaxis using NOACs after acute stroke.
Graduated Compression Stockings
With the controversy over whether the benefits prophylactic anticoagulation outweighed the risk, graduated compression stockings (GCS) became the mainstay of prophylaxis in the UK [96]. GCS, used first in post-operative surgical patients from the 1970s, was adopted for use in stroke patients despite lack of evidence for benefit in stroke. Many physicians, however, felt it to be ineffective in stroke and there was a small trial which was inconclusive [9]. Then the publication of the large CLOTS Trial in 2009, with over 1,200 patients in each group including ICH, showed no significant difference in the incidence of proximal DVT when using thigh-length GCS compared to avoidance of GCS, 10 % vs. 10.5 % OR 0.97 (0.75–1.26). This was a combination of clinical and subclinical DVT diagnosed with compression ultrasound. The incidence of clinical DVT was 2.9 % vs. 3.4 % OR 0.84 (0.53–1.31). There was also no significant difference in the incidence any DVT (proximal and distal) 16.3 % vs. 17.7 % OR 0.9 (0.73–1.11), any clinical DVT 4.4 % vs. 4.8 % OR 0.9 (0.62–1.31), and any PE 1 % vs. 1.6 % OR 0.65 (0.32–1.11). In addition, GCS was associated with a higher incidence of skin breakdown, ulceration, and necrosis 5 % vs. 1 % OR 4.16 (2.40–7.27) [32]. This was a clear result that thigh-length stockings were of no benefit for DVT prophylaxis after stroke. There were few unanswered questions, however, when these results were taken in conjunction with the CLOTS 2 trial, as to why thigh-length stockings were of no benefit and yet better than below-knee stockings. Did below-knee stockings cause DVT or was there a small benefit which could not be shown [97]? Nevertheless, from these trials the recommendation is to avoid any form of compression stocking after stroke and subsequent guidance have reflected this [59, 62].
Intermittent Pneumatic Compression
Like the previous prophylactic measures against VTE, intermittent pneumatic compression (IPC) was first used in surgery and has been shown to prevent DVT (Fig. 6.1). It is thought to prevent DVT by reducing venous stasis, but there is also some conflicting evidence that it may have an effect on fibrinolytic pathways [43]. Its efficacy in prevention of PE was not so clear, as its benefits in non-surgical patients. Guidelines have recommended it use especially in ICH for some time, but there were a few inconclusive trials in acute stroke until the CLOTS 3 Trial, a multicentre trial, published in 2013 [33]. The trial with over 1400 patients in each arm including 13% ICH showed that IPC did reduce the incidence of proximal DVT compared to control by 3.6 % (8.5 % vs. 12.1 %), with adjusted OR 0.65 (0.51–0.84 p = 0.01). There was also a significant reduction in all DVT, clinical and subclinical, proximal and distal 16.2 % vs. 21.1 % OR 0.72 (0.6–0.87 p = 0.01) and in all clinical DVT 4.6 v 6.3 % OR 0.72, (0.52–0.99, p = 0.045). There was no significant difference in the incidence of PE but numbers were small 2.0 % vs. 2.4 %. The combined VTE incidence of any DVT and PE was significantly reduced 17.2 % vs. 22.6 % OR 0.72 (0.59–0.86, p = 0.00035). At 30 days, there was a non-significant reduction in mortality 10.8 % vs. 13.1 % OR 0.8 (0.63–1.01, p = 0.057), but the combined incidence of VTE and death was significantly reduced 27.2 % vs. 34.1 % OR 0.72 (0.61–0.84, p < 0.0001). There was no excess of DVT and PE in the post-treatment period when sleeves were removed, to indicate that IPC simply deferred VTE to later. At 6 months there were a few more DVT and PE and many more deaths, but significant combined difference between the two groups still held 35.6 % vs. 42.3 % (p = 0.002). Using the Cox model adjustments, there was also a significant reduction in the cumulative hazard of death during the 6 months after randomisation in the IPC group, with a hazard ratio of 0.86 (0.74–0.99) p = 0.042. Although there was no difference in confirmed PE, the autopsy rate was low, and the authors suggest that the difference in mortality may be due to a reduction in undiagnosed PE that contributed to death.
Fig. 6.1
Intermittent pneumatic compression (IPC) device with sleeves applied to the patient’s legs
The IPC has a few drawbacks in that the use is not advised with skin conditions like dermatitis and leg ulcers, congestive cardiac failure, and other conditions causing severe oedema and significant peripheral vascular disease, which were all exclusion criteria in the study. The incidence of skin breaks and ulceration with use was higher in the treatment group (3 % vs. 1 %, p = 0.02), though the authors mention that local investigators did not attribute many of these to the IPC, occurring after the sleeves were removed, or on the heels, which are not covered by the sleeves. A more common drawback was the fact that IPC was not tolerated by all patients, and the mean time the devices were on for was 12.5 days, with a median of 9, significantly less than the 30 days envisaged. Perfect adherence was achieved in 31 % of patients, and mean adherence was 59.2 %. Also, from practical experience, the sleeves often had to be taken on and off for therapy, and there are sometimes concerns that if for any reason the sleeves are left off for a reasonable length of time, legs need to checked for DVT and scanning done. The other concern raised is the effect of IPC on other DVT, for instance more proximal DVT in the pelvis and IVC, and that it would clearly have no benefit in preventing upper-extremity DVT. Nevertheless, despite these concerns, this is one intervention that showed unequivocal benefit on reducing DVT and mortality without significant side effects and should be considered first-line for prevention. Some guidelines have it on equivalence with LMWH, but more recent guidelines published have started to make it first-line prevention [60, 61], especially in ICH [70]. On analysis, however, there were no significant cost- or quality-adjusted survival gains [98].
Diagnosis of VTE
Despite prevention measures, post-stroke VTE occur, and even if subclinical VTE are not sought by some form of screening it is known from recent trials that the incidence of symptomatic DVT in the modern era of stroke care will be between 1 and 5 %, and incidence of symptomatic PE will be in the order of 1–3 % [32, 33, 89]. Bearing in mind all the while, there is a much higher subclinical incidence in the order of 20 % for all types of DVT, 10 % for proximal DVT, and perhaps 2–5 % for PE. As mentioned at the start of this chapter, stroke patients may not be able to report limb discomfort, and factors like hemiplegic oedema and pneumonia can mask VTE. So the clinician’s index of suspicion must be high and there should be awareness of subtle signs such as slight increases in calf diameter and tension, a raised respiratory rate and/or heart rate, and minor changes in oxygen saturation. It has to be said these clinical signs may only have moderate predictive value, for instance in the MRDTI study of severe ischemic stroke patients a calf diameter change of 1 cm had a sensitivity of 44 % and positive predictive value of 53 %, whilst using 2 cm change had a greater positive predictive value of 78 % but sensitivity fell to 28 % [99]. Nevertheless, under current processes of care, relying on clinical signs are the mainstay of diagnosis, and knowing the greater risks associated with clinical VTE, it is better to investigate any suspicion and not dismiss them or presume infection or oedema.
The alternative avenue for VTE diagnosis, which has been proposed because of the unreliability of clinical symptoms and signs, would be screening for high-risk patients using a combination of stroke-related features such as severity and immobility and the d-dimer test. As yet, there has not been any recommendation for such a course in clinical guidelines, so now outside of clinical trials, only those suspected of VTE will go on to further investigation.
d-dimer
d-dimer is a cross-linked fibrin degradation product that is generated during thrombus formation and has been used with clinical features to select patient presenting to the emergency departments with a swollen leg or chest pain for further imaging for VTE. In VTE in the general population the plasma level of d-dimer can be increased eight times more than controls and then levels fall with the duration of symptoms and introduction of anticoagulant therapy [100, 101].
The use of d-dimer testing as a screening tool to aid the diagnosis of VTE is less useful in stroke patients, as concentration of d-dimer is increased in conditions associated with enhanced fibrin formation and subsequent degradation by plasmin, such as age, cancer, surgery, infections, acute coronary syndromes, cardiac or renal failure, atrial fibrillation, and stroke [102–105], so standard d-dimer thresholds cannot be relied upon.
Nevertheless, the same d-dimer tests are used in stroke as in other groups of patients, enzyme-linked immunosorbent assays (ELISAs), which have very high sensitivity but low specificity and traditionally take longer to anlyse and quicker modern latex agglutination tests, which tend to be somewhat less sensitive but more specific. More recently, highly sensitive latex agglutination tests have been developed, so that their performance characteristics tend to approach those of the ELISAs [100, 101].
Many studies have shown that d-dimer levels are elevated after stroke and remain elevated for some weeks, but did not select out confounding variables. A paper from the MRDTI study mentioned earlier excluded patients with VTE and also those with other confounders like inter-current infection and found that the median level still remained elevated above standard thresholds. At 2 days post-stroke, the overall median value for the VIDAS ELISA assay was 652 ng/ml (433–1,097) higher than the standard diagnostic threshold of 500 ng/ml. This was even more pronounced in older patients (>70 year) and in patients with severe strokes, with 2-day median of 1,051 ng/ml (663–1,441 ng/ml) and 858 ng/ml (623–1,882 ng/ml). There was no group where standard thresholds would be useful. Median levels in total anterior circulation infarct levels were higher, but lacunar levels were still above the threshold (VIDAS 1,251 ng/ml vs. 721 ng/ml). In those aged <70 year and in non-severe strokes, the median levels of 515 ng/ml (343.5–717 ng/ml) and 612 ng/ml (430–858 ng/ml) were lower but still above the standard [24].
Studies have tried to find a higher d-dimer level which could give greater specificity whilst maintaining the safety of high sensitivity to use for DVT screening and exclusion. A study in the 1990s using an older ELISA assay showed that optimal cut point for predicting DVT with d-dimer was 1,591 ng/ml, resulting in 79 % sensitivity and 78 % specificity, however that would mean missing some DVT, so the group found that lowering the d-dimer cut point to 1,092 ng/ml improved the sensitivity to 100 %, albeit with a fall in specificity to 66 % [106]. Various things, including the older batch assay and the rehabilitation setting with recruitment 3 months after stroke, probably means this d-dimer level is not applicable for use for acute stroke today. The more recent MRDTI study found the VIDAS ELISA test of 2,096 ng/ml and the IL agglutination test of 1,174 ng/ml had a sensitivity of 78 % and 83 % for proximal DVT, but at the cost of imaging 30 % of the patients [99].
Related posts:
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
