Venous thromboembolism (VTE) is a serious and potentially fatal complication of acute ischemic stroke (AIS)8,30,33). The incidence of deep vein thrombosis (DVT) during the first two weeks after stroke ranges from 27% to 75% in untreated patients14). Pulmonary embolism (PE) is the most common cause of death between the first two and four weeks after AIS17). Because DVT is often asymptomatic in AIS patients, it is important to identify risk factors capable of predicting the occurrence of DVT. In previous studies, risk factors for DVT included comorbidities and neurological deficits such as older age, atrial fibrillation, and a higher degree of limb paresis18,22). Additionally, prolonged venous stasis caused by bed rest, cardiac failure, obesity, and dehydration is another risk factor in critically ill patients in the intensive care unit21). Although recent reports have shown that dehydration is a potential risk factor for worse clinical outcomes in AIS, the relationship between dehydration and the risk of venous thromboembolism in AIS has not been clearly identified20,28). The aim of this study was to investigate the relationship between dehydration, as measured by the blood urea nitrogen (BUN)/creatinine (Cr) ratio, and the risk of VTE in patients with AIS.

This retrospective, single-center study included AIS patients who were admitted to our stroke unit between January 1, 2012 and December 31, 2013. The study protocol was approved by the Institutional Review Board of the College of Medicine, The Catholic University of Korea (HC14RISI0035).

AIS was diagnosed based on clinical signs, symptoms, and brain magnetic resonance imaging (MRI) at admission. Information collected at initial assessment included the following: demographic characteristics, stroke severity, vascular risk factors, comorbidities, and baseline laboratory results from blood samples collected on the day of admission (blood cell counts, lipid profile, BUN level, Cr level, serum osmolality [SO], international normalized ratio [INR], erythrocyte sedimentation rate [ESR], and C-reactive protein [CRP] level).

The severity of the patients’ neurological deficits was evaluated using the National Institutes of Health Stroke Scale (NIHSSS) on the day of admission. Stroke severity was defined as moderate (NIHSS score of 8 to 19) or severe (NIHSS score of 20 or more)9). For each patient, the severity of lower extremity weakness and range of motion were reassessed three days after admission in a standardized examination performed by experienced rehabilitation staff. Manual muscle test (MMT) results were classified as follows: grade 5, movement against gravity plus full resistance; grade 4, movement against gravity plus some resistance; grade 3, complete range of motion against gravity with no resistance; grade 2, full or partial range of motion with no gravity; grade 1, slight contractility without any movement; and grade 0, no evidence of contractility (complete paralysis).

The inclusion criteria for this study were as follows: 1) diagnosis of AIS based on clinical presentation and diffusion-weighted MRI; 2) AIS diagnosed within 72 hours of symptom onset; and 3) AIS accompanied by lower extremity weakness. The exclusion criteria were the presence of any of the following: 1) transient ischemic stroke; 2) hemorrhagic stroke; 3) age <18 or >88; 4) current treatment with anticoagulation therapy; 5) active cancer; 6) hospital stay <2 weeks; 7) history of DVT or PE; 8) comorbidities influencing neurological status such as seizures or acute respiratory distress; 9) severe stroke (NISSS score ≥20); 10) immobility before stroke onset; 11) hormone replacement therapy; and 12) history of heart failure, chronic obstructive pulmonary disease, or renal failure.

Early management of AIS was performed according to guideline recommendations9). We administered dual antiplatelet therapy consisting of treatment with aspirin and clopidogrel while the patients were hospitalized. Pharmacologic prophylaxis, such as unfractionated heparin and low molecular weight heparin (LMWH), was not routinely used for VTE prevention. Thromboembolic deterrent stockings were supplied to all patients. Patients diagnosed with VTE were treated with full doses of LMWH. An inferior vena cava filter was also placed if needed.

Patients were divided into two groups: VTE and non-VTE. The VTE group was classified as individuals who were diagnosed with DVT or PE at any time during hospitalization. Diagnosis of VTE was based on clinical and radiological findings. The presence of DVT was confirmed by a lower extremity computed tomography scan, whereas PE was confirmed by computed pulmonary angiography showing the presence of pulmonary emboli or a ventilation-perfusion scan. Diagnostic testing was performed in patients who experienced clinical symptoms suspicious for VTE such as pain, tenderness, swelling, warmth, redness, discoloration, surface vein distention, Homans’ sign or Luck’s sign in one or both legs, cyanosis, chest pain, or respiratory distress.

Of the 182 patients included in the study, 165 (90.7%) were not diagnosed with VTE (non-VTE group) and 17 (9.3%) were diagnosed with VTE. Among the 17 patients who were diagnosed with VTE, two had accompanying DVT and PE. Table 1 lists the baseline clinical and demographic characteristics of both groups. VTE was diagnosed at a mean of 21 days (range, 11–39 days). The patients in the VTE group were older and a higher proportion was female, but the differences between the groups were not statistically significant. Body mass index, history of smoking, and medical history such as diabetes mellitus, hypertension, hyperlipidemia, atrial fibrillation, and stroke were similar in both groups.

The VTE group presented with higher NIHSS scores at admission compared with patients in the non-VTE group; however, these findings were not statistically significant (NIHSS: 6.5 vs. 8.4, p=0.131; NIHSS ≥8: 36% vs. 53%, p=0.163). Laboratory parameters (blood cell counts, lipid profile, BUN level, Cr level, BUN/Cr ratio, SO, SO>295, INR, ESR, and CRP level) on admission were similar in the two groups. However, the proportion of patients with BUN/Cr ratio >15 was significantly different between non-VTE and VTE groups (61% vs. 94%, OR 10.1, 95% CI 1.31–78.31, p=0.007; Fig. 1). Additionally, the VTE group showed more severe lower extremity weakness and more frequent range of motion (ROM) limitation compared with patients in the non-VTE group, although these differences were not significant (weakness grade: 2.5 vs. 1.3, p=0.07; ROM limitation: 53% vs. 64%, p=0.346). However, when a cutoff of grade <3 for lower extremity weakness was used, the non-VTE and VTE groups were significantly different (40% vs. 76%, OR 5, 95% CI 1.56–16, p=0.004; Table 2).

In a regression analysis including relevant confounders and potential predictors of age, sex, stroke severity, low limb weakness, ROM limitation on admission, diabetes, hypertension, hyperlipidemia, atrial fibrillation, smoking, previous stroke, and laboratory findings. Two parameters were identified as predictive factors for VTE: BUN/Cr ratio >15 on admission (OR 8.75, 95% CI 1.11–68.49, p=0.039) and the presence of lower extremity weakness of grade <3 (OR 4.38, 95% CI 1.34–14.26, p=0.014; Table 3).

In this study, VTE was reported in 9.3% of AIS patients with prolonged hospitalization. We compared the demographic and clinical characteristics of 17 patients with VTE and 165 patients without VTE. Multivariate analyses showed that lower limb weakness (poor, trace, or zero using the MMT grading system) and dehydration (defined by BUN/Cr ratio >15) were significant independent risk factors for VTE.

The incidence of VTE in immobilized post-stroke patients ranges from 10% to 75%4). Such a wide variation may be attributed to differences in diagnostic methods and timing, the inclusion of symptomatic or asymptomatic patients or of immobilized or mobile patients, and stroke severity among different studies13,16,24). According to several previous studies, Asians appear to be less prone to development of VTE than other races36). We found that VTE developed in 9.3% of our patients. In a recent large trial (effectiveness of thigh-length graduated compression stockings to reduce the risk of deep vein thrombosis after stroke [CLOTS trial 1]), symptomatic or asymptomatic VTE was found in 10.0% of the patients allocated to use thigh-length graduated compression stockings and in 10.5% of the patients allocated to not use graduated compression stockings; these rates are in agreement with our results10).

Our finding that the risk of VTE is increased in patients with severe lower limb weakness is supported by multiple previous studies. In general, patients who have suffered an ischemic stroke are at increased risk for VTE. Additionally, patients with lower extremity paresis face greater risks of developing VTE because limb paresis and immobilization reduce muscle-pumping activity in the legs, thereby resulting in reduced venous blood flow. Thus, partial or total immobilization due to lower limb paresis in stroke patients is considered to be the greatest risk factor for VTE13). A recent meta-analysis of epidemiological studies found that the risk of VTE in immobilized stroke patients is double that in non-immobilized stroke patients25). Risk factors of VTE in the general population are generally recognized to include older age, obesity, cigarette smoking, hormonal replacement, pregnancy, previous medical illness (such as cancer, COPD, congestive heart failure, thrombophilia, smoking, and hypertension), stroke with limb paresis, and recent surgery. In our study, all AIS patients had limb paresis, and approximately 40% of the patients were unable to resist gravity, confirming the higher risk of VTE among patients with more severe lower limb weakness.

Dehydration is a frequent clinical problem and a marker of poor prognosis in many clinical conditions, including pneumonia and myocardial infarction. Dehydration is also a marker of clinical deterioration in AIS and may be caused by decreased oral water intake due to disturbed consciousness or dysphagia. Although several clinical and laboratory tests have been applied for accurate diagnosis of dehydration severity, there is no definite absolute cutoff value. Weight loss, skin turgor, deep breathing, and capillary filling time have some limitations because baseline levels were not evaluated31,34). Urea concentration ≥100 mg/dL, bicarbonate level ≥15 mEq/L, and serum osmolality have also been used for the diagnosis of dehydration severity11,37). One observational study found an association between elevated osmolality (>296 mOsm/kg) on admission following acute stroke (90% ischemic) and mortality within three months after adjusting for potential confounding factors6). Additionally, dehydration is assumed to predispose patients to VTE. Although there is little evidence indicating dehydration as an independent risk factor for VTE in AIS patients3), dehydration has been suggested to increase the risk of VTE during long-duration and long-distance air travel, a phenomenon known as economy class syndrome2). Dehydration reduces total plasma volume, decreases cardiac output, reduces blood flow to muscle, and can change blood rheology (an increase in local hematocrit and blood viscosity)12). Co-existence of two of three major risk factors of Virchow’s triad (alterations in blood flow, vascular endothelial injury, and alterations in the constitution of the blood) predisposes some dehydrated patients to a heightened risk of VTE35).

In our study, dehydration measured using a biochemical index (BUN/Cr ratio >15) less than three days post-stroke was significantly associated with the development of VTE in univariate and multivariable analyses, with an OR of 8.75 by multivariable analysis, indicating that dehydration is an independent risk factor for the development of VTE in AIS patients. BUN/Cr ratio is one of the common laboratory tests used to evaluate prerenal azotemia, acute tubular necrosis, and dehydration, with a threshold of 201,5,19,26). Recently, several studies have shown that BUN/Cr ratio >15 is a reference marker for early neurological deterioration, three-month outcome, and hospital cost in AIS patients7,20,29). Our study provides added significance of BUN/Cr ratio >15 regarding VTE in AIS, although the mechanisms by which initial hydration status and hydration therapy can influence the incidence of VTE, mortality, or functional recovery in ischemic stroke remain unclear15,23). Fluid replacement may reduce the occurrence of dehydration and increase cerebral blood flow and oxygen delivery, and maintaining systemic blood pressure after acute stroke may contribute to improved clinical outcomes after AIS27,32). Also, given the association between VTE and dehydration after AIS, our findings suggest that encouraging fluid replacement may also reduce the likelihood of developing VTE especially in dehydrated patients with AIS presenting with BUN/Cr ratio >15. No evidence-based method for predicting the occurrence of VTE after stroke currently exists, and our findings suggest that a high Bun/Cr ratio (BUN/Cr >15) in the acute stage is an independent predictor of VTE in AIS patients. Although there is no single ‘gold standard’ clinical or laboratory measure of the hydration status, BUN and Cr levels are routinely measured in all medical fields. Our study thus provides a new perspective on predicting the occurrence of VTE in AIS patients, and this cutoff may apply to all stroke patients. Our findings support the use of prophylactic anticoagulation if the Bun/Cr ratio is elevated regardless of the AIS mechanism and stroke severity.

Although our findings are preliminary, based on our results, physicians may consider maintaining hydration in previously dehydrated patients with severe lower limb weakness in the early stage of ischemic stroke in order to reduce the chance of VTE. Treatment with LMWH is the preferred method for preventing VTE regardless of the subtype of ischemic stroke. A larger-scale randomized controlled trial is required to replicate the findings of this preliminary study.

There are several limitations of this study: 1) we cannot draw strong conclusions because the sample size was small and the control group was retrospectively selected; 2) there was a lack of uniformity in the thromboprophylaxis administered to the patients; 3) we excluded patients with severe stroke (NIHSS ≥20) or a short hospital stay, regardless of the reason; 4) we did not perform routine diagnostic procedures such as ultrasonography to identify cases of VTE and thus might have missed clinically silent cases; 5) we were not able to perform autopsies on patients who died during the study period; and 6) only serum markers were used to measure dehydration.

In conclusion, our study shows that severe lower limb paresis is associated with VTE. Moreover, an elevated serum BUN/Cr ratio (>15) is independently associated with an increased risk of VTE. Therefore, physicians should be aware of the danger of dehydration and be attentive to patients with higher BUN/Cr ratios and severe leg paresis. Further investigations of the pathophysiology of VTE are needed, and a larger controlled study is required to assess whether these results hold true in a larger patient population.