Sage Journals: Discover world-class research

Abstract

This study developed and evaluated a tailored nomogram to predict the potential occurrence of early lower extremity deep vein thrombosis (LDVT) in patients receiving thrombolytic therapy. We performed several logistic analyses on the training cohort and created a corresponding nomogram to forecast early LDVT. The classification accuracy and the accuracy of predicted probabilities of the multiple logistic regression model were evaluated using area under the curve (AUC) and the calibration graph method. According to the multivariate logistic regression model homocysteine, previous history of hypertension and atrial fibrillation, indirect bilirubin, age, and sex was identified as independent determinants of early LDVT. The nomogram was constructed using these variables. The calibration plots showed a good agreement between the predicted and observed LDVT possibilities in the training and validation cohorts with AUCs being 0.833 (95% CI: 0.774-0.892) and 0.907 (95% CI: 0.801-1.000), respectively. Our nomogram offers clinicians a tool for predicting the individual risk of LDVT in the early stage of acute ischemic stroke in patients receiving thrombolytic therapy, which could lead to early intervention.

Keywords

nomogram stroke thrombolytic therapy deep vein thrombosis‌

Introduction

Cerebrovascular diseases, the most common of which is stroke, primarily affect middle-aged and older adults and cause significant harm to both their physical and mental fitness. These diseases pose a considerable financial burden and cause significant distress, both to the patients and their families, and have considerable socioeconomic implications. Globally, stroke is the third leading cause of death. The risk of death and disability due to cerebrovascular conditions is ever increasing, leaving approximately two-thirds of survivors with varying degrees of paralysis and symptomatic disabilities. According to the literature, older age, limb paresis, and atrial fibrillation (AF) are the major risk factors of post-stroke lower extremity deep vein thrombosis (LDVT).¹ LDVT is common in the acute stage of ischemic stroke. Previous studies have reported that the prevalence of LDVT after stroke ranges from 20% to 70%, depending on the modality used to detect its presence. Previous studies have shown that the incidence of LDVT varies by ethnicity and race, with Asian stroke patients having a lower risk than expected (4.8%-45%).² The prevalence of immobilized post-stroke patients ranges from 10% to 75%, depending on the diagnostic method and time of evaluation.¹ There have been some studies on the incidence of LDVT in stroke patients; however, only a few studies predicting the risk factors for complications associated with patients receiving thrombolytic therapy have been conducted. Thus, our study developed and evaluated a tailored nomogram to predict the potential occurrence of early LDVT in patients receiving thrombolytic therapy. To the best of our knowledge, our study is thus the first to fully elucidate the relationship between early LDVT and thrombolysis and create a predictive model.

The continuous development of diagnostic and therapeutic methods, such as thrombolytic therapy, has improved the rehabilitation efforts and recovery time of patients in the acute stage of stroke.³ The first-line recanalization treatment for rescuing ischemic brain tissue is limited to thrombolytic therapies; the intravenous administration of recombinant tissue-type plasminogen activator (rt-PA) may effectively stimulate fibrinogen within the fibrinolytic system to optimize cerebral thrombi cleavage.^4,5 Therefore, thrombolytic treatment could be considered a preventative measure against LDVT⁶; however, one study found that in 21% of patients treated with rt-PA thrombolytic therapy did not prevent LDVT, this prevalence is almost as high as that in stroke patients who did not receive thrombolytic therapy.⁷ In addition, a study conducted by Mori et al found that 17.6% of patients with acute ischemic stroke (AIS) receive thrombolytic therapy.⁸ According to previous reports, there is barely a difference in the occurrence of LDVT between AIS patients who received thrombolytic therapy and those who did not.⁶ The following mechanisms may explain why there were no variations noted between patients who underwent cerebral venous thrombolysis and those who did not: (1) In the initial stages following thrombolysis, patients must have a bed rest, which can result in sluggish and stagnant venous blood flow to the paralyzed lower extremities; (2) Dehydrating agents that contribute to blood in a hypercoagulable state may be used to prevent cerebral edema and cranial hypertension; (3) The development of LDVT may result from this therapeutic increase in coagulation factor activity.

A nomogram is a visual calculation tool that is easy to read, straightforward, and practical, it displays the results of multifactorial regression analysis using a collection of disconnected lines on the same platform. Classified as a quantitative prediction tool, it has grown in popularity and usefulness in the clinical field.^9,10 Approximately 60% of clinicians are well-versed in the nomogram, 56% of healthcare workers are accustomed to reference lists, and 21% of clinicians prefer chart formats such as decision trees.¹¹ In response to the daily utility aspect of these 3 tools, 74% of individuals found the format of the nomogram, which is the most frequently applied tool, to be good.¹¹ It is useful for estimating output risk at the economic, preventative, and individual levels and can be used in clinical settings to improve prediction accuracy. A nomogram can be fully integrated with constrained clinical statistics to concentrate on tailored risk predictions and provide various diagnostic and therapeutic approaches.¹² Only a very few studies investigating the risk factors for LDVT in stroke thrombolysis patients have been conducted, fewer still have modeled them as a logistic regression model for risk prediction. This study establishes and validates a risk prediction model for LDVT after thrombolysis in patients with AIS. The nomogram is primarily based on univariate analysis and multifactorial logistic regression analysis to provide a foundation and reference for medical prevention and intervention of post-thrombolysis in patients with AIS.

Materials and Methods

Study Design and Participants

In this study, 377 patients with AIS who received thrombolytic therapy at our medical facility were included between January 2016 and August 2022. There were 177 patients with LDVT in group 1 and 200 patients without LDVT in group 2. The inclusion criteria were as follows: (1) an MRI scan verified the AIS diagnosis, which also satisfied the diagnostic standards outlined in the AHA/ASA Guidelines for stroke¹³; (2) the patient's general and clinical information was complete (age ≥ 18 years; no sex restrictions); (3) stroke began within 4.5 h and was treated with rt-PA intravenous thrombolytic therapy; (4) ultrasound evaluation during hospitalization; and (5) Baseline information collected from patients or their families. The exclusion criteria were as follows: (1) patients with lower extremity varicose veins or a history of LDVT; (2) patients with significant limb movement limitations caused by previous cerebral infarction; (3) patients with a history of fracture within the last 3 years; (4) patients who had been hospitalized in our hospital for longer than 10 days from the time of stroke onset.

Diagnosis of LDVT

LDVT is characterized by erythema or pitting edema of the skin, tenderness, and superficial branching veins.¹⁴ Cramping is a typical LDVT sensation in the lower extremity. LDVT should be suspected if a patient exhibits any of the abovementioned clinical symptoms for an unknown period of time.¹⁵ However, because clinical signs and symptoms can be highly diverse and unique to each individual, the prognosis of LDVT that is based solely on medical presentation is unreliable. Therefore, imaging tests such as vascular ultrasound can validate or exclude prediction.¹⁴ In this study, patients regularly underwent ultrasounds of the arteries and veins in each of their lower extremities within the first 3 days of hospitalization. A pressure probe shows a limited indentation of the venous lumen, color Doppler shows a flow-filling defect or no detectable flow, and ultrasound of the veins in the lower limbs reveals uneven echoes populating the lumen. LDVT is diagnosed when any of the abovementioned conditions are satisfied.

Data Collection

We examined information on individuals’ medical histories, laboratory test findings, and NIHSS scores. The patients provided their underlying medical and pharmaceutical histories. Our laboratory department performed more than 20 other indices in addition to the coagulation index, thyroid function, blood routine, and liver and kidney functions. After thrombolytic therapy, all biochemical indicators were chosen from the data within 24 h of admission. The NIHSS score was scored within 30 min after thrombolysis.

Statistical Analysis

All continuous data are presented as medians with interquartile ranges or as means ± standard deviations, as appropriate, whereas categorical values are presented as relative frequencies and proportions. Both categorical and continuous variables were used to compare baseline information. In R4.2.1, the training and validation sets were randomly grouped at a 3:1 ratio using the car (https://socialsciences.mcmaster.ca/jfox/Books/Companion/) and survival (https://CRAN.R-project.org/package=survival) package. The following data analysis was performed in the training set: Continuous variables were first tested for normality and then analyzed using the independent samples t-test or the Mann-Whitney U test. Categorical variables were analyzed using the χ² test. These analyses were conducted using IBM SPSS statistics (Version 26.0), and variables with P values ≤0.05 were considered meaningful and were retained in a collinearity check; variance inflation factors (VIFs) larger than 10 were rejected. Finally, the remaining variables were included in the multivariate logistic regression, and variables with P values <0.15 were selected into the R software (R4.2.1 statistical software program). The rms package (https://CRAN.R-project.org/package=rms) was used to establish a model for LDVT formation in patients with AIS. An evaluation model was established based on the regression analysis results. The receiver operating characteristic (ROC) curve reflects how accurately the model classifies individuals into the occurrence and nonoccurrence categories. The extent to which the calibration chart response model predicts the agreement between the predicted outcome in an individual and the true possibilities.¹⁴ Next, the data from the validation set were introduced into the calibration and ROC curves to verify the feasibility.

Results

Patient Characteristics

From January 2016 to August 2022, 377 AIS patients treated in our hospital were enrolled in this analysis; using R4.2.1 statistical software, 283 patients of the selected populations were categorized into the training cohort. The remaining 94 patients were categorized into the validation cohort. We used the IBM SPSS statistics (Version 26.0) software to determine that there were no appreciable changes (P > 0.05) between the training and validation cohorts in terms of fundamental patient characteristics or laboratory data (Table 1). There was no statistical difference in the incidence rates of LDVT between the 2 cohorts after thrombolytic therapy: 129 of 283 (45.6%) patients in the training cohort developed LDVT, and 48 of 94 (51.1%) participants in the validation cohort were developed LDVT.

Table 1.

Baseline Characteristics of AIS Thrombolysis Patients in the Training Cohort and Validation Cohort.

Variables		Training_cohort(n = 283)	Validation_cohort(n = 94)	P value
LDVT, n(%)	No	154 (54.4%)	46 (48.9%)	.422
LDVT, n(%)	Yes	129 (45.6%)	48 (51.1%)	.422
Laboratory parameters
LYM (×10⁹)		1.56 ± 0.76	1.63 ± 0.8	.466
ALB (g/L)		37.57 ± 3.83	37.4 ± 3.89	.839
UA (μmol/L)		316.55 ± 113.7	323.7 ± 125.02	.743
HCY (μmol/L)		12.19 ± 6.42	12.44 ± 7.46	.775
WBC (×10⁹)		8.72 ± 3.38	8.65 ± 3.07	.691
Neutrophon (×10⁹)		6.56 ± 2.99	6.52 ± 3.01	.665
Monocytes (×10⁹)		0.54 ± 0.25	0.55 ± 0.21	.433
RBC (×10¹²)		4.38 ± 0.56	4.31 ± 0.56	.312
HB (g/L)		134.31 ± 17.83	130.97 ± 18.41	.129
PLT (×10⁹)		212.31 ± 63.91	213.69 ± 64.42	.983
TP (g/L)		66.17 ± 5.74	65.59 ± 5.55	.397
ALT (U/L)		19.6 ± 13.8	23.26 ± 29.03	.255
AST (U/L)		25.42 ± 12.49	29.87 ± 34.55	.233
STB (μmol/L)		15.7 ± 6.95	15.8 ± 7.71	.624
IB (μmol/L)		10.39 ± 5	10.22 ± 5.19	.478
DB (μmol/L)		5.53 ± 2.58	5.66 ± 3.04	.756
Urea (μmol/L)		5.42 ± 2.27	6.01 ± 4.05	.067
Cr (μmol/L)		70.74 ± 23	73.99 ± 33.49	.525
Urea.Cr		19.54 ± 6.33	20.21± 6.5	.323
HDL (mmol/L)		1.17 ± 0.33	1.14 ± 0.27	.752
CK (U/L)		130.91 ± 184.75	112.47 ± 87.82	.961
CK-MB (ng/mL)		13.27 ± 6.62	16.31 ± 23.81	.647
Na (mmol/L)		139.77 ± 3.01	140.1 ± 3.01	.623
K (mmol/L)		3.74 ± 0.36	3.76 ± 0.36	.560
TH (nmol/L)		106.6 ± 22.22	106.77 ± 18.23	.841
TSH (mIU/L)		1.69 ± 5.61	2.67 ± 12.08	.850
FT3 (pmol/L)		4.47 ± 1.12	4.41 ± 0.63	.793
FT4 (pmol/L)		11.53 ± 2.35	11.73 ± 2.06	.570
PT (s)		14.38 ± 1.66	14.22 ± 1.22	.384
INR		1.13 ± 0.18	1.11 ± 0.12	.253
APTT (s)		37.63 ± 5.16	36.79 ± 5.1	.324
FIB (g/L)		2.78 ± 0.94	2.81 ± 0.99	.473
DD (mg/L)		3.34 ± 3.64	3.57 ± 4.15	.791
CRP(mg/L)		17.02 ± 31.14	11.88 ± 15.31	.965
Demographic characteristics
Age (year)		69.18 ± 12.03	68.5 ± 12.43	.611
Gender, n(%)	Woman	117 (41.3%)	44 (46.8%)	.419
Gender, n(%)	Man	166 (58.7%)	50 (53.2%)	.419
NIHSS		8.35 ± 6.26	8.3 ± 6.01	.932
Past history
HBP, n (%)	No	55 (19.4%)	21 (22.3%)	.645
HBP, n (%)	Yes	228 (80.6%)	73 (77.7%)	.645
AF, n (%)	No	208 (73.5%)	65 (69.1%)	.494
AF, n (%)	Yes	75 (26.5%)	29 (30.9%)	.494
Smoke, n (%)	No	189 (66.8%)	65 (69.1%)	.767
Smoke, n (%)	Yes	94 (33.2%)	29 (30.9%)	.767
alcohol, n (%)	No	202 (71.4%)	66 (70.2%)	.933
alcohol, n (%)	Yes	81 (28.6%)	28 (29.8%)	.933

Abbreviations: AF, atrial fibrillation; AIS, acute ischemic stroke; ALB, albumin; ALT, alanine aminotransferase; APTT, activated partial thromboplastin time; AST, aspartate aminotransferase; CK, creatine kinase; CK-MB, creatine kinase isoenzymes; Cr, creatinine; DB, direct bilirubin; DD, D-dimer; FIB, fibrinogen; fT3, serum free triiodothyronine; fT4, serum free thyroxine; HB, hemoglobin; HBP, hypertension; HCY, homocysteine; HDL, high-density lipoprotein; IB, indirect bilirubin; INR, international normalized ratio; K, potassium ion; LYM, lymphocyte count; Na, sodium ion; NIHSS, National Institute of Health Stroke Scale; PLT, platelets; PT, prothrombin time; RBC, red blood cells; STB, total serum bilirubin; TH, thyroid hormone; TP, total protein; TSH, thyroid-stimulating hormone; UA, uric acid; WBC, white blood cells.

Independent Predictors of LDVT in AIS Patients With Thrombolytic Therapy

We divided the patients in the training cohort into 2 individual groups, with and without LDVT; the baseline characteristics of the subgroups are presented in Table 2. Univariate analysis of the training cohort reported that there were differences between the 2 groups in lymphocyte count (LYM, P = 0.001), homocysteine (HCY, P = 0.000), monocytes (P = 0.023), red blood cells (RBCs, P = 0.000), hemoglobin (HB, P = 0.006), platelets (PLT, P = 0.043), indirect bilirubin (IB, P = 0.003), plasminogen time (PT, P = 0.050), D-dimer (DD, P = 0.005), age (P = 0.000), sex(P = 0 .010), history of hypertension (HBP, P = 0.000), AF (P = 0.052), and smoking habits (P = 0.013).

Table 2.

Univariate and Multivariate Analysis of the Baseline Characteristics in the Training Cohort.

Variables	Univariate analysis		Multivariate analysis
	Statistic	P value	VIF	β	P value	OR	95% CI
	Statistic	P value	VIF	β	P value	OR	Lower limit	Upper limit
Laboratory parameters
LYM (×10⁹)	Z = −3.338	.001	1.256	−0.633	.189	0.531	0.206	1.366
UA (mmol/L)	Z = −1.781	.075
HCY(mmol/L)	Z = −4.151	.000	1.127	0.150	.055	1.161	0.997	1.353
WBC (×10⁹)	Z = −0.751	.452
Neutrophon (×10⁹)	Z = −0.400	.689
Monocytes (×10⁹)	Z = −2.280	.023	1.215	−0.574	.609	0.563	0.062	5.090
RBC (×10¹²)	T = −4.230	.000	10.048
HB (g/L)	T = −2.784	.006	10.301
PLT (×10⁹)	Z = −2.019	.043	1.203	−0.003	.517	0.997	0.988	1.006
TP (g/L)	T = −0.653	.514
ALT (U/L)	Z = −1.201	.230
AST (U/L)	Z = −0.550	.582
STB (mmol/L)	Z = −1.387	.165
IB (mmol/L)	Z = −2.979	.003	1.253	0.071	.122	1.074	0.981	1.175
DB (mmol/L)	Z = −0.505	.614
Urea (mmol/L)	Z = −0.706	.480
Cr (mmol/L)	Z = −0.971	.332
Urea.Cr	Z = −0.068	.945
HDL (mmol/L)	Z = −0.016	.987
CK (U/L)	Z = −1.331	.183
CK-MB (ng/mL)	Z = −1.919	.055
Na (mmol/L)	T = 0.577	.564
K (mmol/L)	T = −1.755	.080
TH (nmol/L)	T = 1.751	.081
TSH (mIU/L)	Z = −1.384	.166
fT3 (pmol/L)	Z = −1.525	.127
Ft4 (pmol/L)	T = 1.052	.294
PT (s)	T = −1.970	.050	2.168	−0.078	.613	0.925	0.685	1.250
INR	T = −1.531	.127
APTT (s)	T = −1.039	.300
FIB (g/L)	Z = −0.999	.318
DD (mg/L)	Z = −2.823	.005	1.211	0.065	.491	1.067	0.888	1.282
CRP (mg/L)	T = −0.051	.712
Demographic characteristics
Age (year)	T = 4.060	.000	1.429	0.076	.005	1.079	1.023	1.137
Gender, n (%)	χ² = 6.685	.010	1.858	−0.984	.131	0.374	0.104	1.339
NIHSS	Z = −0.948	.343
Past history
HBP, n (%)	χ² = 20.278	.000	1.167	−1.228	.087	0.293	0.072	1.195
AF, n (%)	χ² = 3.778	.052	1.193	−0.969	.099	0.380	0.120	1.201
Smoke, n (%)	χ² = 6.228	.013	1.596	0.092	.887	1.097	0.307	3.920
Alcohol, n (%)	χ² = 0.000	.984

Abbreviations: AF, atrial fibrillation; VIF, variance inflation factor; LYM, lymphocyte count; HCY, homocysteine; RBC, red blood cell; HB, hemoglobin; PLT, platelets; IB, indirect bilirubin.

In the training cohort, indicators with P values less than 0.05 and VIF <10 were substituted for further multifactor logistic regression analysis. The following correlation indicators <0.15 were finally obtained.

As a measurement of multicollinearity in a multiple linear regression model, the VIF represents the ratio of the variance of the estimated regression coefficients to the variance when no linear correlation is assumed between the independent variables. The possibility of collinearity among independent variables increased with the VIF value. A VIF of >10 is usually identified as an indicator of severe multicollinearity in a regression model. In the covariance analysis, factors with differences in the training set described above were included, and those with a VIF >10 (RBC and HB) were excluded.

The training cohort's age, sex, history of HBP and AF, and laboratory indicators of LYM, HCY, monocytes, PLT, IB, PT, and DD were associated with a higher risk of developing early LDVT after thrombolytic therapy, according to univariate analysis and VIF (Table 2). These filtered components were added to the multivariate regression analysis to choose the independent predictors of LDVT. The final results of the multivariate logistic regression analysis revealed that these 6 variables (age, sex, HBP, AF, IB, and HCY) acted as independent predictors of LDVT among AIS patients who received thrombolytic therapy (Table 2).

Construction of the Nomogram for LDVT in AIS Patients With Thrombolytic Therapy

A nomogram about LDVT risk in patients receiving thrombolytic therapy was constructed using R software based on the 6 independent predictors of LDVT (Figure 1). The figure is pictorial, where each variable is listed separately, with a corresponding number of points assigned to a particular magnitude of the variable.¹⁶ Each indicator vertically corresponds to the bottom score line with the corresponding value. Next, the points from each indicator value were summed. The sum is located at the Total Points Scale and is vertically projected onto the bottom axis; thus, we can obtain an individualized assessment of LDVT following thrombolytic therapy.

Figure 1.

Nomogram for the possibility of early LDVT in AIS patients receiving thrombolytic therapy. In sex, 1 represents man, and 0 represents woman. In the history of HBP and AF, 1 represents yes, and 0 represents no. The corresponding score for each indicator is vertically down, and the sum can be obtained on the total points scale for each patient. Finally, the possibility of LDVT is the risk of “LDVT” vertically corresponding to “Total Points.” AF, atrial fibrillation; AIS, acute ischemic stroke; IB, indirect bilirubin; HBP, hypertension; HCY, homocysteine; LDVT, lower extremity deep vein thrombosis.

Feasibility Validation of the Nomogram

Finally, the nomogram was validated using a training cohort. The ROC curves are shown in Figure 2, and the prediction accuracy of the nomogram is shown in Figure 3. The area under the curve (AUC) value represents the ability to identify patients with different outcome events. The nomogram exhibited a relatively good discriminative capacity based on an AUC value of 0.833 (95% CI: 0.774-0.892; Figure 2A). Additionally, calibration plots revealed acceptable agreement between the predicted and actual LDVT probabilities, showing that the predicted risk of LDVT in the multivariate logistic regression model was closely approximated by actual observations (Figure 3A). In the training cohort, the mean absolute error between the prediction probability and actual probability was 0.021. Verification was also assessed in the validation cohort by comparing the predicted probability of the nomogram and the actual probability of each patient. The calibration plot showed a mean absolute error of 0.021 between the predicted and actual probabilities, indicating good consistency (Figure 3B). The AUC of the predictive nomogram was 0.907 (95% CI: 0.801-1.000; Figure 2B).

Figure 2.

The ROC curves of the nomogram for estimation of lower extremity deep vein thrombosis (LDVT) among the training cohort (A) and validation cohort (B). (A) The AUC value is 0.833 (CI: 0.774-0.892). (B) The AUC value is 0.907 (CI: 0.801-1.000). AUC, area under the curve; CI, confidence interval; ROC, receiver operating characteristics curves.

Figure 3.

The calibration plots for training cohort (A) and validation cohort (B). (A) Mean absolute error = 0.021 (training cohort); (B) Mean absolute error = 0.020 (validation cohort), indicating that in an individual, the agreement between the predicted outcome and the true probability is not bad.

Discussion

AIS is a rapidly evolving illness with high rates of disability and mortality. Intravenous thrombolytic therapy is currently the first-line treatment for AIS and is highly valued by medical professionals and researchers alike because of its proven stability and added value. However, patients who accept thrombolytic therapy are still at risk of developing LDVT, either as a result of slow, stagnant venous blood flow in the paralyzed lower limbs; as a result of thrombolytic agents stimulating coagulation pathways; or as a result of hypercoagulable blood because of high cranial pressure throughout subsequent remedies such as a dehydrating agent. The risk factors for developing LDVT during hospitalization have only been vaguely analyzed for patients who have received thrombolytic therapy. Thus, to enhance the prognosis of LDVT within these groups, this study examines the LDVT risk factors in AIS patients treated with intravenous thrombolysis in recent years and creates a nomogram to investigate the risk factors for LDVT after intravenous thrombolysis and their impact on clinical prognosis.

This study established and validated an easy-to-use and intuitive score for predicting the risk of early LDVT in AIS patients receiving thrombolytic therapy. It applies to patients who were treated within 10 days of stroke onset who do not have a previous LDVT history. From the model, we can deduce that age, IB, HCY, and female sex are negative factors associated with the thrombosis process. Furthermore, with a history of AF and HBP, the male sex is a protective factor for LDVT. Good consistency was observed in both the training and validation cohorts.

Three fundamental factors can be identified as the etiology of LDVT from a pathophysiological point of view: venous wall damage, blood with hypercoagulability, or specifically sluggish blood flow.¹⁷ Our model explanation revolves around these 3 elements.

In our investigation, older age increased the likelihood of developing LDVT in patients receiving thrombolytic therapy. This could be due to the following reasons: (1) older adult patients have features of osteoporosis which can make walking difficult, reduce exercise capacity, and decrease muscle mass, which results in the reduced pumping function of muscles, contributing to a slow and stagnant blood flow condition, thus promoting thrombosis¹⁸; (2) aging is known to decrease the elasticity and smoothness of the vascular wall of the intima, furthermore, due to aging insufficiency of closure occurs via venous valve atrophy, and the coagulation system is activated by damaged endothelial cells, all of which lead to a highly solidified blood state¹⁹; (3) through aerobic respiration or inflammation in hepatocytes and macrophages under physiological conditions, signaling molecules called reactive oxygen species (ROS) are formed, albeit in small amounts. The natural aging process occurs because of ROS negative function in cell differentiation and apoptosis. Free radicals can be generated by UV radiation exposure as people age and by poor eating habits, high levels of stress, vigorous activity, and other unexpected situations. The probability of intima damage and blood clotting is enhanced with the age-related ascension of oxygen free radicals and ROS.²⁰

The risk of LDVT appears to differ between the sexes, with women scoring higher than men, as found through the nomogram. A woman's life is considered to have frequent fluctuations in prothrombotic activity. In their youth, pregnancy, menstrual cycles, and oral contraceptives are common, after which menopause is inevitable, and hormone replacement therapy may be used, all of which can have potential effects on the development of cardiovascular disease, including LDVT.²¹ In addition, through the research conducted by Towfighi and Kotseva et al, middle-aged women start to have a temporal trend of an overall increase in the prevalence of cardiovascular disease, while a decrease in prevalence in men is mainly due to an increase in smoking habits, diabetes, and hypertension in the former.^22,23

Patients with hypertension are more susceptible to increased pulse pressure. The reactive congestive index, systemic perfusion, and increased blood flow rate are positively correlated with increased pulse pressure, which can reduce blood stasis. In contrast, such an effect cannot be achieved with decreased blood pressure. However, most patients with hypertension in this study with a history of disease onset before admission were mostly treated with antihypertensive therapy and had good blood pressure control. The above analysis explains that a history of treated hypertension, similar to normal blood pressure, is a risk factor for LDVT development in patients with AIS receiving thrombolysis.

Whether new or not, patients with AF have received anticoagulant therapy such as heparin or vitamin K antagonists and new oral anticoagulants such as dabigatran and rivaroxaban. Thanks to the timely use of anticoagulants, compared with stroke patients without AF and thus received antiplatelet therapy, patients receiving anticoagulant therapy are less viscous in blood and at a lower risk of developing thrombosis.²⁴ Therefore, this may explain why a history of AF is a protective factor against the development of LDVT in AIS patients receiving thrombolysis.

The metabolic end-product of heme is IB, catalyzed by heme oxygenase, and has neuroprotective and neurotoxic effects. Within the normal physiological range, higher levels of IB can inhibit the oxidation of linoleic acid and phospholipids, scavenge free radicals, enhance endothelial antioxidant action, and reduce cardiovascular disease incidence.²⁵ However, interestingly, IB, more than physiological doses, is cytotoxic and harmful and can damage mitochondrial function by increasing mitochondrial membrane permeability. Therefore, the pathological levels of IB may indicate a poor prognosis.²⁶

Homocysteine is positively associated with venous thromboembolism formation; we explain that HCY is involved in vascular endothelial damage and the formation of hypercoagulable blood,²⁷ which may both be a pathogenic agent of arterial wall lesions and a thrombogenic factor.²⁸

Previous studies confirmed that increased CRP levels are a risk factor for LDVT.²⁹ However, our study showed that the CRP index was not statistically significant in univariate logistic regression analysis. This variable was collected within 24 h after the patient's thrombolytic treatment, the value of this variable differs at different time point measurements, the magnitude of change varies from person to person, and a predictor measured at fixed time points cannot reflect the dynamic changes in inflammation over a prolonged period.

Based on the multifactor analysis, the nomogram integrates multiple predictors to achieve individualized and accurate prediction of the probability of an event. In our investigation, age, IB, HCY, sex, history of AF, and HBP are all factors associated with the development of LDVT in stroke patients receiving thrombolytic therapy. Based on these results, we created a predicting model with good AUC values for both sensitivity and specificity, showing that the model is of high clinical relevance. Several models are currently available to predict LDVT in stroke patients, such as the Liu, IMPROVE, and Padua scores in clinical work. None of these studies specifically targeted a specific group of stroke patients receiving thrombolytic therapy. Our nomogram targets this specific population with high accuracy and good identification, allowing for early detection and diagnosis, which can improve clinical prognosis.

Limitations

Our study has some limitations. First, because of the inherent nature of retrospective studies, the sequence of acute cerebral ischemic events and the occurrence of LDVT cannot be fully determined. Second, our hospitalized patients underwent lower extremity ultrasonography within the first 3 days of admission, and there was no follow-up after discharge. Therefore, it was unavailable to accommodate patients who developed LDVT within a relatively short time after acute stroke events. Third, the data sources for our nomogram were based only on retrospective analysis of a single-center database. Thus, to validate the scalability of the final model in other centers, we need to conduct a prospective study to confirm its reliability further.

Conclusions

In this study, HCY and IB, the only 2 laboratory indicators are routine clinical tests. The risk of LDVT development in patients receiving thrombolytic therapy can be reasonably predicted by combining the patients’ basic conditions, such as age, sex, and history of AF. Our nomogram, which can be easily promoted and applied, offers clinicians a tool for predicting the individual risk of LDVT development in the early stage of AIS in patients undergoing thrombolytic therapy, which could lead to early intervention.

Footnotes

Authors’ Note

This study was approved by the Ethics Committee of the First Affiliated Hospital of Wenzhou Medical University (Acceptance Number KY2023-R040) and was performed in accordance with the latest version of the Declaration of Helsinki promulgated by the National Institute of Health. This was a retrospective study and all data were anonymized. Due to the retrospective nature of this study patient informed consent was waived. The First Affiliated Hospital of Wenzhou Medical University's Ethics Committee approved the waiver of informed consent. Due to privacy and ethical restrictions data presented in this study are not publicly available. Data can be made available on reasonable request from the corresponding author.

Author Contributions

Yuyan Wang is the only first author. Yuyan Wang contributed to conceptualization, methodology, software, formal analysis, writing—original draft preparation; Yiling Sun contributed to visualization; YiKai Wang and RuJun Jin contributed to validation; Bei Shao contributed to conceptualization, investigation, funding acquisition; Wanli Zhang contributed to resources, supervision, writing—review and editing; Xuan Liu and Miaokao Cao contributed to data curation. All authors have read and agreed to the published version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

Bei Shao

References

Bembenek

Karlinski

Kobayashi

Czlonkowska

. Early stroke-related deep venous thrombosis: risk factors and influence on outcome. J Thromb Thrombolysis. 2011;32:96‐102.

Kim

Heo

Chang

Kim

. Prediction of deep vein thrombosis by ultrasonography and D-dimer in Asian patients with ischemic stroke. BMC Neurol. 2020;20:257.

Rinde

Smabrekke

Mathiesen

, et al. Ischemic stroke and risk of venous thromboembolism in the general population: the Tromso study. J Am Heart Assoc. 2016;1:1‐8.

Berge

Whiteley

Audebert

, et al. European Stroke Organisation (ESO) guidelines on intravenous thrombolysis for acute ischaemic stroke. Eur Stroke J. 2021;6:I–LXII.

Phipps

Cronin

. Management of acute ischemic stroke. Br Med J. 2020;368:l6983.

Liu

Guo

, et al. Incidence and risk factors of lower-extremity deep vein thrombosis after thrombolysis among patients with acute ischemic stroke. Pharmgenomics Pers Med. 2021;14:1107‐1114.

Balogun

Roberts

Patel

Pathansali

Kalra

Arya

. Clinical and laboratory predictors of deep vein thrombosis after acute stroke. Thromb Res. 2016;142:33‐39.

Mori

Yoshioka

Tanno

. Frequency of deep vein thrombosis at admission for acute stroke and associated factors: a cross-sectional study. Thromb J. 2021;19:62.

Bai

Liu

Xiong

, et al. Nomogram to predict overall survival after gallbladder cancer resection in China. World J Gastroenterol. 2018;24:5167‐5178.

10.

Wang

Song

Zhou

Liu

Zhang

. From past to future: bibliometric analysis of global research productivity on nomogram (2000-2021). Front Public Health. 2022;10:997713.

11.

Capitanio

Jeldres

Shariat

Karakiewicz

. Clinicians are most familiar with nomograms and rate their clinical usefulness highest, look-up tables are second best. Eur Urol. 2008;54:958‐959.

12.

Caulfield

Menezes

Marignol

Poole

. Nomograms are key decision-making tools in prostate cancer radiation therapy. Urol Oncol. 2018;36:283‐292.

13.

Powers

Rabinstein

Ackerson

, et al. Guidelines for the early management of patients with acute ischemic stroke: 2019 update to the 2018 guidelines for the early management of acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2019;50:e344‐e418.

14.

Di Nisio

van Es

Büller

. Deep vein thrombosis and pulmonary embolism. Lancet. 2016;388:3060‐3073.

15.

Streiff

Agnelli

Connors

, et al. Guidance for the treatment of deep vein thrombosis and pulmonary embolism. J Thromb Thrombolysis. 2016;41:32‐67.

16.

Balachandran

Gonen

Smith

DeMatteo

. Nomograms in oncology: more than meets the eye. Lancet Oncol. 2015;16:e173‐e180.

17.

Wenger

Sebastian

Engelberger

Kucher

Spirk

. Pulmonary embolism and deep vein thrombosis: similar but different. Thromb Res. 2021;206:88‐98.

18.

Sadeghian

Divsalar

Fadil

Tavakolian

Blaber

. Canadian aging and inactivity study: Spaceflight-inspired exercises during head-down tilt bedrest blunted reductions in muscle-pump but not cardiac baroreflex in older persons. Front Physiol. 2022;13:943630.

19.

Brunelli

. Deep vein thrombosis/pulmonary embolism: prophylaxis, diagnosis, and management. Thorac Surg Clin. 2012;22:25‐28. v.

20.

Jakubczyk

Dec

Kałduńska

Janda

. Reactive oxygen species – sources, functions, oxidative damage.

21.

Renda

Patti

Lang

, et al. Thrombotic and hemorrhagic burden in women: gender-related issues in the response to antithrombotic therapies. Int J Cardiol. 2019;286:198‐207.

22.

Ovbiagele

ATLZB

. 2009. Sex-Specific Trends in Midlife Coronary Heart Disease Risk and Prevalence.

23.

Kotseva

Wood

De Backer

, et al. Cardiovascular prevention guidelines in daily practice: a comparison of EUROASPIRE I, II, and III surveys in eight European countries. Lancet. 2009;373:929‐940.

24.

Seiffge

Werring

Paciaroni

, et al. Timing of anticoagulation after recent ischaemic stroke in patients with atrial fibrillation. Lancet Neurol. 2019;18:117‐126.

25.

Maruhashi

Kihara

Higashi

. Bilirubin and endothelial function. J Atheroscler Thromb. 2019;26:688‐696.

26.

Wang

Zhong

. Serum bilirubin and ischaemic stroke: a review of literature. Stroke Vasc Neurol. 2020;5:198‐204.

27.

Khandanpour

Willis

Meyer

, et al. Peripheral arterial disease and methylenetetrahydrofolate reductase (MTHFR) C677 T mutations: a case-control study and meta-analysis. J Vasc Surg. 2009;49:711‐718.

28.

Bal Dit Sollier

Drouet

. Involvement of thrombophilia in coronary thrombosis. Ann Cardiol Angeiol (Paris). 2017;66:365‐372.

29.

Bembenek

Karlinski

Kobayashi

Czlonkowska

. Deep venous thrombosis in acute stroke patients. Clin Appl Thromb Hemost. 2012;18:258‐264.

Nomogram Prediction for Lower Extremity Deep Vein Thrombosis in Acute Ischemic Stroke Patients Receiving Thrombolytic Therapy

Abstract

Keywords

Introduction

Materials and Methods

Study Design and Participants

Diagnosis of LDVT

Data Collection

Statistical Analysis

Results

Patient Characteristics

Independent Predictors of LDVT in AIS Patients With Thrombolytic Therapy

Construction of the Nomogram for LDVT in AIS Patients With Thrombolytic Therapy

Feasibility Validation of the Nomogram

Discussion

Limitations

Conclusions

Footnotes

Authors’ Note

Author Contributions

Declaration of Conflicting Interests

Funding

ORCID iD

References